0% found this document useful (0 votes)

77 views354 pages

Methods in Analytical Political Theory Compress

This book serves as a comprehensive guide to key methods in analytical political theory, including contractualism, reflective equilibrium, and positive political theory. It emphasizes practical application by detailing the problems each method addresses, the underlying principles, and concrete practices for implementation. The aim is to make methodological ideas accessible to students and contribute innovatively to research methods in the field.

Uploaded by

Aabid

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

77 views354 pages

Methods in Analytical Political Theory Compress

Uploaded by

Aabid

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 354

Methods in Analytical Political Theory

This is the ﬁrst book to explain how to use key methods in analytical
political theory. The methods discussed include contractualism, reﬂec-
tive equilibrium, positive political theory, thought experiments and
ideological analysis. Many discussions of political theory methods
describe and justify these methods with little or no discussion of their
application, emphasizing ‘what is’ and ‘why do’ over ‘how to’. This book
covers all three. Each chapter explains what kinds of problems in poli-
tical theory might require researchers to use a particular method, the
basic principles behind the method being proposed, and an analysis of
how to apply it, including concrete principles of good practice. The book
thus summarizes methodological ideas, grouped in one place and made
accessible to students, and it makes innovative contributions to research
methods in analytical political theory.
Methods in Analytical Political
Theory

Edited By

Adrian Blau
King’s College, London
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
4843/24, 2nd Floor, Ansari Road, Daryaganj, Delhi – 110002, India
79 Anson Road, #06–04/06, Singapore 079906

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781107098794
DOI: 10.1017/9781316162576
© Edited By Adrian Blau 2017
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2017
Printed in the United Kingdom by Clays, St Ives plc
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Names: Blau, Adrian, 1972– editor.
Title: Methods in analytical political theory / edited by Adrian Blau, King’s
College, London.
Description: New York, NY : Cambridge University Press, 2017.
Identiﬁers: LCCN 2016047820 | ISBN 9781107098794 (hardback)
Subjects: LCSH: Political science – Methodology. | Political science – Research –
Methodology.
Classiﬁcation: LCC JA71 .M4225 2017 | DDC 320.01–dc23
LC record available at https://lccn.loc.gov/2016047820
ISBN 978-1-107-09879-4 Hardback
ISBN 978-1-107-49170-0 Paperback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party Internet Web sites referred to in this publication
and does not guarantee that any content on such Web sites is, or will remain,
accurate or appropriate.
Methods in Analytical Political Theory

Edited By

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

Edited By

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

List of Contributors page vii

Preface ix

1 Introduction: A ‘How-to’ Approach 1

adrian blau
2 How to Write Analytical Political Theory 18
robert e. goodin
3 Thought Experiments 21
kimberley brownlee and zofia stemplowska
4 Reﬂective Equilibrium 46
carl knight
5 Contractualism 65
jonathan quong
6 Moral Sentimentalism 91
michael l. frazer
7 Realism 112
robert jubb
8 Realistic Idealism 131
david schmidtz
9 Conceptual Analysis 153
johan olsthoorn
10 Positive Political Theory 192
alan hamlin
11 Rational Choice Theory 217
brian kogelmann and gerald gaus

v
vi Contents

12 Interpreting Texts 243

adrian blau
13 Comparative Political Thought 270
brooke ackerly and rochana bajpai
14 Ideological Analysis 297
jonathan leader maynard
15 How to Do a Political Theory PhD 325
keith dowding and robert e. goodin

Index 327
Contributors

Brooke Ackerly is Associate Professor in the Department of Political

Science at Vanderbilt University.
Rochana Bajpai is Senior Lecturer in Politics in the Department of
Politics and International Studies at SOAS.
Adrian Blau is Senior Lecturer in Politics in the Department of Political
Economy at King’s College London.
Kimberley Brownlee is Associate Professor of Legal and Moral
Philosophy in the Department of Philosophy at the University of
Warwick.
Keith Dowding is Professor of Political Science in the School of Politics
and International Relations at the Australian National University.
Michael L. Frazer is Lecturer in Political and Social Theory in the
School of Politics, Philosophy, Language and Communication
Studies at the University of East Anglia.
Gerald Gaus is the James E. Rogers Professor of Philosophy in the
Philosophy Department at the University of Arizona.
Robert E. Goodin is Distinguished Professor in the School of
Philosophy at the Australian National University.
Alan Hamlin is Emeritus Professor of Political Theory in Politics,
School of Social Sciences at the University of Manchester, and Senior
Research Fellow in the Department of Political Economy at King’s
College London.
Robert Jubb is Lecturer in Political Theory in the Department of Politics
and International Relations at the University of Reading.
Carl Knight is Lecturer in Political Theory in the School of Social and
Political Sciences at the University of Glasgow.

vii
viii List of Contributors

Brian Kogelmann is a PhD student in the Philosophy Department at

the University of Arizona.
Jonathan Leader Maynard is Departmental Lecturer in International
Relations in the Department of Politics and International Relations at
the University of Oxford.
Johan Olsthoorn is a post-doctoral fellow of the Research Foundation
Flanders (FWO)/KU Leuven.
Jonathan Quong is Associate Professor of Philosophy in the School of
Philosophy at the University of Southern California.
David Schmidtz is Kendrick Professor of Philosophy (College of
Social & Behavior Sciences) and Eller Chair of Service-Dominant
Logic (College of Management) at the University of Arizona.
Zoﬁa Stemplowska is Associate Professor of Political Theory at
Worcester College, Oxford.
Preface

I had the idea for this book in late 2008. The central concept has stayed
the same – a ‘how-to’ book with practical advice in bold type. But much
else has changed. I owe a great deal to early guidance from my former
colleagues at the Manchester Centre for Political Theory (MANCEPT),
especially Alan Hamlin, Jon Quong and Hillel Steiner. My Cambridge
University Press editor, John Haslam, and my anonymous reviewers
had invaluable advice about content and scope: they made this book far
better.
I have greatly enjoyed working with my contributors. I wanted every
chapter to give clear, concrete advice grounded in examples of good or
bad practice. It was a pleasure to see this take shape and I hope our
readers will benefit.
I dedicate this book to the memory of my father, Nat Blau, who passed
away in 2010 after a long and productive life. He was a doctor, researcher
and teacher. He would have approved of a book that seeks to be helpful,
practical and well exemplified.
My father had a strong sense of our intellectual limits: for example, he
sometimes said that ‘if a theory explains all the facts, the theory must be
wrong, because some of the facts are wrong.’ This book is written not in
the expectation that our arguments and tips are right, but in the hope that
enough of them are right that we can advance the debate – and that
readers’ challenges to our arguments and tips will do so further.
This is the first book of its kind. It must not be the last.

ix
1 Introduction: A ‘How-to’ Approach

Adrian Blau

1 Introduction
This is the first ‘how-to’ handbook in political theory. It describes
different methods (‘what is method X?’), justifies them (‘why do method
X?’), and explains what to do (‘how should I do method X?’). ‘How-to’
guidance is our main aim. Political scientists have hundreds of such
handbooks, from general overviews on research design to textbooks on
specific techniques. Philosophers have dozens of handbooks on logic and
critical thinking. But political theorists have no handbook on how to apply
the methods we use.
Existing books on political theory methods typically describe and
justify methods without much detail on implementation – ‘what is’ and
‘why do’ more than ‘how to’. For example, Quentin Skinner’s book on
methods in the history of political thought contains relatively little prac-
tical guidance (Skinner 2002: 40–2, 75–6, 79–80, 114–20). If you want to
interpret texts like Skinner, you will learn more from his actual research.
Similarly, there is little advice on how to implement the methods and
approaches covered in the fine volume edited by Leopold and Stears
(2008).
When we praise or criticize work in political theory, ‘how-to’ principles
are often implicit. For example, if you find that some definitions are
clearer than others, you have already grasped some principles of concep-
tual clarity, consciously or subconsciously. But you will struggle to find
much published guidance on defining terms clearly. Our book aims to
make such implicit principles explicit and adds new how-to principles as
well.
Much of our focus, then, is on the logic of inference – on how best to draw
robust conclusions, on how to justify our conclusions against actual or
potential critics. ‘It came to me in a dream’ is not typically considered
good methodology in political theory. Rather better is ‘I carefully distin-
guished freedom and autonomy, used thought experiments to test their
relative importance, and engaged with comparative political thought to

1
2 Chapter 1

see if these ideas ﬁt non-Western cultures.’ True, some scholars will

disagree with methods like thought experiments, just as some social
scientists reject statistical analysis or ethnography, say: healthy disciplines
see disputes about methods. But there are principles of good practice in
statistical analysis and ethnography for social scientists who want to use
these methods, and principles of good practice in thought experiments for
political theorists who want to use this method. Each chapter in this book
outlines such principles.
Our book’s key contribution, then, is practical guidance on what to do
and what to avoid in the methods you may use in political theory. Since
your methods affect your answers, how-to guidance should help.

2 What Is the ‘How-to’ Approach?

Each author has placed ‘how-to advice’ – sometimes obvious, sometimes
not – in bold type. Some chapters spread the advice through the text;
sometimes it is more concentrated. After reading a chapter, you can flick
through it looking only at the advice in bold and remind yourself how to
apply that method.
Consider the method of thought experiments, where we imagine situa-
tions that help us probe moral or political problems. Is medical experi-
mentation on humans ethically wrong? We could try to answer this by
seeing how we react to the idea of Nazis experimenting on people in
concentration camps. But this particular example will probably bias us
against medical experimentation. To answer the question more reliably,
we should consider medical experimentation in less extreme cases, with-
out Nazis. ‘Be sensitive to possible narrative-framing biases’ is Brownlee
and Stemplowska’s advice in the thought experiments chapter. This may
seem obvious, but many published thought experiments have such biases.
Learning to spot this will improve your own reasoning and help you
criticize some existing arguments.
A second example comes from the textual interpretation chapter,
where I write: ‘indicate how confident you are in your interpretations.’
We can never know for sure what Marx meant by ‘class’ or why
Machiavelli wrote The Prince. If two explanations are plausible but
one is better supported by the evidence, we do not help our readers by
pretending that one is definitely right while the other is undeniably
wrong. This advice may seem obvious, but such indications of
uncertainty are not common in published research on textual
interpretation.
We are not presenting a neutral handbook: all of our prescriptions are
contestable. Indeed, we want them to be contested. Being explicit about
Introduction: A ‘How-to’ Approach 3

how-to principles will hopefully clarify the issues, stimulate debate, and
advance research in political theory.

3 Why Use a ‘How-to’ Approach?

There are five main reasons for a ‘how-to’ approach. First, and most
mundane, students and academics often have to discuss methods in
dissertations, PhD proposals, funding applications and so on. A better
developed methods literature will help us compete with other researchers.
You can explain to readers if you are using consent contractualism, fair-
ness contractualism or rationality contractualism; conceptual analysis as
resolution, extensional analysis and/or disambiguation; normative, his-
torical, interpretive and/or critical comparative political thought; and
so on.
Second, ‘how-to’ analysis helps us answer ‘why do’ questions: under-
standing how to apply a method can help us decide whether to use it, in at
least three ways. One way is where we avoid a method because we wrongly
think that it cannot be implemented. For example, realist political theory
has often been criticized as overly negative, lacking a constructive pro-
gramme. By showing – at last! – how to do realist political theory, Jubb’s
chapter in this volume strengthens the case for realism.
Another way in which ‘how to’ addresses ‘why do’ is where we wrongly
think that a given approach is not how we do things and thus not what we
should do. My chapter on textual interpretation partly targets historians
who see philosophical analysis as something for political theorists and
philosophers. But when we see how we actually interpret texts, we find
that everyone does philosophical analysis: historians should thus learn
how to do it well. Hamlin’s chapter argues that since many of us do
positive political theory without realizing it, we should recognize this,
and do it better. Frazer’s chapter on moral sentimentalism sees the
abstract, rationalistic nature of much political theory as a fairly recent
invention. Great political theorists like Hume and Smith often took
sentimentalist approaches, as does Rawls in some respects: sentimental-
ism is not alien to political theory.
But the most important way in which ‘how to’ helps us answer ‘why
do’ is where people wrongly reject a method in general due to parti-
cular applications of it: such criticisms can evaporate if we see how to
apply the method better. Examples in this book include some objec-
tions to rational choice theory discussed by Kogelmann and Gaus,
some challenges to reflective equilibrium answered by Knight, and
some doubts about thought experiments considered by Brownlee
and Stemplowska.
4 Chapter 1

Third, how-to guidance helps us avoid mistakes. Olsthoorn’s chapter

on conceptual analysis notes that some arguments try to make normative
points by definitional sleight-of-hand. It is better to define concepts more
neutrally and argue for the normative position separately. Schmidtz’s
chapter on realistic idealism criticizes act-utilitarians for being curiously
inattentive to consequences, focusing on utility maximization without
considering unintended consequences. Too many political theorists, he
suggests, emphasize thought experiments at the expense of empirical
research, wish away the problem with overly idealized depictions of
human interactions, or recommend massive state powers without con-
sidering corruption. Kogelmann and Gaus note that rational choice ana-
lysis can be undermined by not distinguishing between parametric and
strategic situations.
Fourth, and closely related, how-to guidance can strengthen our argu-
ments. Goodin’s chapter gives many tips for writing and structuring
papers. Brownlee and Stemplowska explain how thought experiments
help us test abstract principles. Ackerly and Bajpai show comparative
political theorists the value of looking beyond elites and beyond texts.
Leader Maynard notes that ideological analysis may help us grasp the
real-world effects of some normative principles.
Fifth, how-to analysis can help us improve methods and thus contribute
to the four points just raised. This is a key aim of our book, which both
summarizes good practice and contributes to these techniques. Knight
rejects Rawls’s view that the key judgements in reflective equilibrium are
judgements held with confidence. Quong distinguishes three types of
contractualism in terms of five dimensions of contractualism. Ackerly
and Bajpai support all four types of comparative political thought, against
those who want to restrict it to one type. My chapter argues that previous
methodological discussions of textual interpretation have largely over-
looked the principles of good practice that actually drive good research.

4 Historical Reﬂections
Of course, debates over methods have always been important in political
theory. Plato’s Republic can be read as highlighting the weaknesses of
Socrates’s method and the strengths of Plato’s new method, ‘dialectic’
(Reeve 1988: 4–9, 21–3). Machiavelli scorns abstract Platonic theorizing:
we should ‘concentrate on what really happens’, because a prince who
follows ‘theories or speculations’ will ‘undermine his power rather than
maintain it’ (Machiavelli 1988: ch. 15, 54). In other words, empirical
observation gives us more guidance than theoretical conjecture about
what works and what does not.
Introduction: A ‘How-to’ Approach 5

Hobbes uses as little empirical observation as possible to derive norma-

tive principles. If we get the methods right, he argues, we will all agree
about politics. The key is reason, which for Hobbes means applying
deductive logic to clearly defined concepts (Hobbes 1991: 4.4, 25–6;
4.9–13, 26–9; 5, 31–7). This shows, for example, that monarchies and
republics have the same amount of liberty (1991: 21.8–9, 149–50).
Hobbes mocks the prevailing methods of political argument, such as
quoting classical authors to support claims (1991: 5.22, 37; 25.12, 180).
Poor methods lead to insufficient authority, thinks Hobbes (1990:
2–4), or to excess authority, for Bentham and Mill (Bentham 1843: 13,
495; Mill 1989: 1.4–7, 8–12; 3.13, 66–7). All three writers dismiss
intuitionism – justifying normative claims by consulting our instincts,
feelings and non-inferential reactions to normative problems (Hobbes
1994: 6.8, 42; Bentham 1996: 2.11–16, 21–31; Mill 1989: 4.12, 84).
Rousseau disagrees: morally pure men can look inside themselves and see
what is good. Conscience is a better philosopher than reason, he writes in
Emile, and we only need reason when conscience fails us (Rousseau 1979:
book 4, 286, 289–90). This from the man who told a recently bereaved
woman that she was luckier than Rousseau: she had only lost her hus-
band, but poor Rousseau had also lost her friendship (Cranston 1997:
87–8). We are lucky that we are not as morally pure as Rousseau.
Although Rousseau defends moral intuition, his political arguments are
not justified by appeal to intuition. Indeed, the draft of The Social Contract
treats intuition as inappropriate in modern society (Rousseau 1997:
1.2.6–14, 154–7). Rousseau’s political method is broadly contractarian,
alongside Hobbes, Locke, Kant, Rawls and Habermas: a normative posi-
tion is justified if all specified agents do or could accept it under certain
conditions (Darwall 2003). The precise version of this idea varies mark-
edly among these writers, but the general method is strikingly different to
that used by writers like Burke, who opposes abstract reasoning and
justifies political actions through prudence, necessity, expedience and
experience (e.g. Burke 1999: 228, 237–8, 280–1). But Burkean justifica-
tion is now rare in academic political theory.
The twentieth century saw considerable innovation and progress in
methods (Wolff 2013). Many of these developments underpin the chap-
ters in this book. Strikingly, though, much political theory still seeks what
Plato and his predecessors sought: conceptual clarity. What does
X mean? How do X and Y differ? Analysis of equality, for example, has
changed hugely in the past twenty-five years because of new distinctions –
between equality, sufficiency and priority, between deontic and telic
egalitarianism, and so on. Many people who saw themselves as favouring
equality now realize that they favour something else. This has not solved
6 Chapter 1

the normative debates, but it has changed the picture and led to better
answers.
This brief historical overview does not imply that good methods ensure
universally accepted answers or that progress is impossible without
a how-to handbook. But methods have clearly been important in political
theory from the beginning, and a how-to handbook could thus bring this
key issue into sharper focus and foster further development.

5 The Scope of This Book

I should brieﬂy discuss the key terms in this book’s title. I will also touch
on what and whom this book has left out.
Obviously, this book’s title is somewhat misleading: a more accurate
title would have been Methods, Methodologies, Techniques and Approaches
in Analytical Political Theory (To The Extent That We Can and Should
Distinguish Analytical and Continental Political Theory Anyway). But even
if that ﬁtted on the front cover, no one would read the book. Mill,
criticizing the ‘utter unreadableness’ of the late Bentham, wrote percep-
tively that Bentham ‘could not bear, for the sake of clearness and the
reader’s ease, to say, as ordinary men are content to do, a little more than
the truth in one sentence, and correct it in the next’ (Mill 1974: 114–15).
I have said a little less than the truth in the title, and corrected it here.
I will not specify what ‘methods’ are, although some chapters do
diverge somewhat from mainstream understandings. For Quong, con-
tractualism is a method and not, as some people think, a substantive
position like utilitarianism or deontology. My chapter on interpreting
texts argues that supposedly different methods in history of political
thought are not different methods: scholars have thus overlooked some
core principles that apparently different methods share. My chapter is
thus more about methodology (seen as the logic of inference) than meth-
ods; Frazer’s chapter, in contrast, is more about an approach and its
techniques than a method and methodology.
This book treats ‘political theory’ very broadly. In effect, we cover
political philosophy, moral philosophy, normative jurisprudence, positive
political theory, history of political thought and more. But we concentrate
on ‘analytical’ political theory, not ‘continental’ political theory. This
distinction is questionable, of course, but roughly, we can associate
analytical political theory with the work of such writers as G. A. Cohen,
Ronald Dworkin, Hobbes, Frances Kamm, Mill, Martha Nussbaum,
John Rawls, Quentin Skinner and Jeremy Waldron, and continental
political theory with the work of such writers as Theodor Adorno,
Judith Butler, William Connolly, Michel Foucault, Hegel, Nietzsche,
Introduction: A ‘How-to’ Approach 7

Richard Rorty and Slavoj Žižek. There is huge variety within these camps,
of course; some writers, like Jürgen Habermas and Bernard Williams,
arguably fit in both camps; and not all of these names would classify
themselves like this. Nor do all authors in this book.
Miller and Dagger (2003: 446–9) outline five principles of analytical
political theory: (a) it is essentially separate from deep metaphysical ques-
tions about the meaning of human life, (b) it involves conceptual clarity
and argumentative rigour, (c) it is normative, (d) it addresses a plurality
of competing values, and (e) it ‘aims to serve as the public philosophy of
a society of free and equal citizens who have choices to make about how
their society will be organised.’
I do not believe we need principles (d) and (e). More importantly,
principle (c) only entails normative political theory, whereas this book is
wider. What did Locke mean by ‘man’? Why did Tocqueville write what
he wrote? These questions are essentially empirical. Was Hayek
a conservative? Does Arendt have a positive conception of liberty?
These questions are both empirical and conceptual: we infer Hayek’s
and Arendt’s beliefs, then compare them to criteria of conservatism and
positive liberty, respectively. What is equality? Do liberty and power
overlap? These questions are fundamentally conceptual. Do Hobbes’s
laws of nature follow necessarily from his premises? Would Rawls’s
principles of justice differ if we changed the motivations of people in the
original position? These questions could be normative, but are in essence
exercises in logic and conceptual clarity. I see no reason to exclude these
questions from analytical political theory.
But ultimately, it does not matter whether you agree with this view of
analytical political theory. Nor does it matter whether you see yourself in
the analytical and/or continental camp, or even as something different. All
that matters is whether some of this book’s guidance helps you.
Although I sought a wide coverage for this book, there are gaps.
We have nothing on international political theory, for example, or on
how political theorists should use empirical research in history and social
science. Such gaps will be criticized, and rightly so. I have some excuses
here, but not for the paucity of women authors – only four out of seven-
teen. When I started inviting authors, it was men who tended to spring to
my mind and to the minds of most of the men and women I sought advice
from. Only late in the day did I see the effects of this implicit bias. Around
the same time I was struck by a perceptive comment on the Feminist
Philosophers blog, which notes that while it might seem arbitrary to invite
a woman to speak at a conference just because she is a woman, it is also
arbitrary that many of us tend to think of men when we are asked who the
leading figures in a field are. I’ve tried to change my thinking and my
8 Chapter 1

actions after reading the Feminist Philosophers blog, the Being a Woman in
Philosophy blog, and the edited book Women in Philosophy (Hutchison and
Jenkins 2013). But this came too late to have much effect on the line-up of
authors. Mea culpa.
So the range of authors is not representative; we have not covered all of
political theory, or even all of analytical political theory; and the title is
a bit of a lie. But what we have is still an exciting venture that seeks to do
something new and important: the ﬁrst book that seeks to explain how to
do much of political theory.

6 Overview of This Book

Robert Goodin’s short introduction on How to Write Analytical
Political Theory offers concise, practical tips on writing clear and effec-
tive analytical political theory. Goodin gives advice on such issues as
structure (e.g. organization = argument), techniques (e.g. distinctions =
arguments) and the process of writing and revising (e.g. don’t overreact to
advice).
Kimberley Brownlee and Zofia Stemplowska’s chapter on Thought
Experiments covers imaginary scenarios such as John Rawls’s original
position and Nozick’s experience machine. For thought experiments to
be useful, the chapter argues, they must be philosophically respectable
and argumentatively relevant. We need not avoid ‘crazy’ or ‘wacky’
thought experiments: indeed, if we understand how-to issues properly,
we see that wacky examples can be beneficial, for example by highlighting
the limits of our conceptual and argumentative structures.
Carl Knight’s chapter on Reflective Equilibrium explains and devel-
ops John Rawls’s influential approach to moral justification, by which we
try to reach an accord between our principles and our judgements. Knight
rejects Rawls’s view that judgements should be confident: Knight prefers
judgements to be considered. Knight also challenges Singer’s objection
that some judgements are emotive, evolutionary responses: this is some-
thing to feed into the reflective equilibrium process, rather, to see if some
judgements are unreliable. Knight offers a step-by-step guide to the
process of reflective equilibrium: we should list the main contending
principles, test those principles with cases, revise principles, and so on.
Knight exemplifies this with an example from his own work, showing how
he came to support a combination of luck egalitarianism and
prioritarianism.
Jonathan Quong’s chapter on Contractualism denies that contractu-
alism is a specific position, like utilitarianism or deontology. Rather, it is
a method by which we consider what idealized agents would or could
Introduction: A ‘How-to’ Approach 9

accept under certain speciﬁed constraints. Or to be precise, contractual-

ism is a family of three methods. Consent contractualism asks what
people would freely consent to, even if we cannot ask them (as when
considering the coercive authority of the state). Fairness contractualism
asks what is fair, even if people would not consent to it (as when card
players discover during a game that a card is missing: since some players
will be doing better than others, we might mistrust their actual, self-
interested views about what to do). Rationality contractualism asks
what is instrumentally rational for people to do, whether or not they
consent, whether or not it is fair. Each type of contractualism involves
different answers to five key questions, such as what agents’ motives and
interests are. Much confusion comes from defenders and critics not
adequately distinguishing these different contractualisms.
Michael Frazer’s chapter on Moral Sentimentalism argues against
the overly rationalistic nature of most modern political theory – cold,
abstract and with too many trolleys. Yet there is ample evidence that
our normative judgements rest in part on emotions. Frazer thus argues
that we should draw not just on reason, but also non-rational faculties
of the human mind, including emotion, imagination and the imagina-
tive sharing of emotion through sympathy or empathy. He looks to
philosophers like Hume and Smith, who engaged their readers with
stories and examples. Moral sentimentalists should replace technical
terminology with the evocative language of everyday life. They should
move from empathetic assessment of particular cases to more general
principles, not vice versa. And they should beware their own biases by
considering multiple perspectives: your assessment of alleged police
injustice will be stronger if you have considered the perspective of the
police, not just the policed.
Robert Jubb’s chapter on Realism discusses the realist critique of ideal
theory. Realists like Bernard Williams criticize Rawls and others for being
overly reliant on general moral claims that neglect the distinctive char-
acter of the political. Politics should come first: realists urge political
philosophers to pay more attention to how political institutions and actors
produce or fail to produce a particular set of goods in response to
a particular set of problems, and to consider how this should constrain
their theorizing. Jubb disagrees with Williams that only liberalism can
adequately answer our basic legitimation demands: various solutions are
possible, and realists should not fall into the trap of themselves being
overly moralistic by appealing to values whose contestation is actually
a key part of the problem. Realists should also engage with particular
political circumstances, developing a finer sense of real political possibi-
lities and motivations.
10 Chapter 1

David Schmidtz’s chapter on Realistic Idealism also criticizes Rawls

for starting with justice. The ﬁrst virtue of social institutions, rather, is to
enable peaceful cooperation. Schmidtz’s criticisms of ideal theory are
different to Jubb’s but complementary. Schmidtz criticizes ideal theorists
who assume away the problems they try to solve. An ideal solution is
one that works with actual agents, not one that would work with idealized
people: we should focus not on a kingdom of ends, but a kingdom of
players. While the political theorist’s job is to some extent to say how the
world should be in the grand scheme of things, actual governance is the
art of compromise in a world that is not a blank canvas. Realistic idealists
should thus focus less on abstract thought experiments and engage more
with history, political economy and empirical insights about the beha-
viour of actual human beings.
Johan Olsthoorn’s chapter on Conceptual Analysis starts, as one
would expect, with careful conceptual analysis of ‘conceptual analysis’.
Olsthoorn distinguishes between concepts and terms, between concepts
and conceptions, between principles and criteria of application, between
different types of concepts and between different types of conceptual
analysis. He offers advice on deﬁning and naming concepts, and on
comparing/contrasting interpretations of concepts advanced by different
writers. Olsthoorn also highlights the limitations of what conceptual
analysis can contribute to normative argument. While conceptual analysis
helps us clarify ideas, make distinctions, and keep separate ideas apart,
disagreements over competing normative ideas are not usually resolvable
by conceptual analysis alone.
Alan Hamlin’s chapter on Positive Political Theory treats positive
political theory as a kind of model, like a map of the London
Underground – abstracting, simplifying and idealizing aspects of the
real world in order to highlight and systematize key features of
a process, institution or argument. For example, in considering whether
voting should be compulsory, we might make assumptions about low
turnout leading to unequal turnout, leading to differential policy impact
and so on. Being clear about such assumptions, even if not in the pub-
lished version of an analysis, can help us think through the relevant steps.
In particular, we can use ‘sensitivity analysis’ to test the relative impor-
tance of various factors, changing one factor at a time to see how it affects
our conclusions. Would compulsory voting be more or less desirable if,
say, unequal turnout did not affect policy? Since many of us do this kind
of thing informally, Hamlin recommends that we should learn to do it
well.
Brian Kogelmann and Gerald Gaus’s chapter on Rational Choice
Theory explains why political theorists should take rational choice theory
Introduction: A ‘How-to’ Approach 11

seriously, whether as a central feature of their work (e.g. game-theoretic

approaches to morality) or to help with other issues (e.g. democratic
theorists interested in institutional design). Part of the chapter involves
explaining core features of rational choice theory, such as the relational
nature of preferences, assumptions such as transitivity and the idea of
Nash equilibria. The chapter deflects common criticisms of rational
choice theory, such as the objection that people aren’t rational, or the
view that rational choice theory isn’t sufficiently predictive. Normatively,
Kogelmann and Gaus discuss applications of rational choice theory in
contractarian arguments about what people would agree to. Importantly,
such deliberations need not exclude considerations of morality or fair-
ness. Still, we should be careful not to overstate what rational choice
theory can offer normative analysis, the chapter argues.
My chapter on Interpreting Texts challenges the dominant cate-
gorization of interpretive approaches, e.g. contextualism, rational
reconstruction, Marxism and Straussianism. These categories have
their uses, but are problematic. For example, contextualism, or the
‘Cambridge School’ approach, is not best seen as a method. The focus
should be on evidence. Contextualists need textual, contextual, moti-
vational and philosophical evidence – as do all of us, ideally. Likewise,
‘reconstruction’ is not a technique just for philosophers, but is some-
thing we all do to resolve ambiguities and gaps in communication.
Unfortunately, the emphasis on different approaches means that meth-
odologists have not outlined core principles of good practice that apply
to all of us (and that already underpin much research). My chapter
collates those principles.
Brooke Ackerly and Rochana Bajpai’s chapter on Comparative
Political Thought sees two key features in comparative political thought
(CPT): an emphasis on bringing African, Chinese, Indian, Islamic and
Latin American ethical-political thought to Western audiences, and an
effort to make political theory more relevant by combining it with political
science, anthropology and history. Ackerly and Bajpai’s stance is plural-
ist, in contrast to the restrictive scope of some CPT scholars. Four main
types of CPT are identified: normative-analytic, historical, interpretive
and critical. These are not necessarily alternatives: scholars may want to
‘triangulate’ different approaches. Ackerly and Bajpai thus stress the
value of a question-driven approach: we should not be too wedded to
a particular approach, but should let the problem guide our methods and
tools.
Jonathan Leader Maynard’s chapter on Ideological Analysis considers
how the study of real-world systems of political ideas enhances analytical
political theory, including normative political theory. For example,
12 Chapter 1

normative theorists can use ideological analysis to reveal the problematic

ideologies that some institutions generate, and vice versa, and can probe the
background ideological assumptions behind normative arguments.
In terms of analysing ideologies themselves, the highly interdisciplinary
nature of ideological analysis makes a cross-disciplinary approach desirable.
Leader Maynard shows how to uncover the content of ideologies, how to
explain ideological attachment and how to analyse the effects of ideologies.
He encourages us not to see ideologies in pejorative terms, as many have;
ideologies are simply the normative and factual ideas that shape our under-
standings and actions. There is no such thing as ‘beyond ideology’.
Robert Goodin and Keith Dowding’s short piece on How to
Do a Political Theory PhD gives advice on how to choose a PhD
topic/problem, and how considering why other scholars are wrong helps
one develop answers. Goodin and Dowding encourage PhD students to
break down the thesis into manageable chunks, but to be ﬂexible: struc-
ture and arguments often change. Planning ahead is especially important
for PhD students considering academic careers: thinking about publica-
tions early on will help.

7 How Do the Chapters Fit Together?

Which chapters complement or compete with each other? I would say –
some would disagree – that reﬂective equilibrium (Knight) is fundamen-
tal to normative political theory, even if we all fall short of the ideal to
greater or lesser extents. Within reﬂective equilibrium, we can use
thought experiments (Brownlee and Stemplowska), contractualism
(Quong), moral sentimentalism (Frazer), realism (Jubb) and realistic
idealism (Schmidtz). Contractualism typically involves thought experi-
ments, but moral sentimentalism, realism and realistic individualism all
oppose some uses of thought experiments, without necessarily rejecting
them entirely. Realists and realistic idealists may have qualms about some
contractualist arguments.
Also fundamental is conceptual analysis (Olsthoorn). Everyone uses
concepts: they are the building blocks of clear thinking, even for primarily
non-normative political theorists.
Positive political theory (Hamlin) is more widespread within political
theory than one may think: we all make assumptions about connections
between different aspects of processes, institutions and arguments.
Rational choice theory (Kogelmann and Gaus) is a subset of positive
political theory.
Interpreting texts (Blau) is not just for people studying history of
political thought: we can easily misread Rawls, for example, by dipping
Introduction: A ‘How-to’ Approach 13

in and quoting ideas out of context. Textual interpretation is particularly

important for comparative political thought (Ackerly and Bajpai) and
ideological analysis (Leader Maynard). But both of these latter chapters
discuss aspects of textual interpretation that range far beyond what my
chapter covers – and both chapters have important lessons for normative
political theorists using reﬂective equilibrium, thought experiments, con-
tractualism, moral sentimentalism, realism, realistic idealism and con-
ceptual analysis.
Finally, the short chapters at the start and end of this book, on how to
do analytical political theory (Goodin) and how to do a political theory
PhD (Goodin and Dowding), will help everyone, especially PhD stu-
dents, in some core skills of political theory and of scholarly practice
more generally.

8 Conclusion: General Lessons from this Book

Some useful general principles can be discerned in this book. One is that
distinctions are vital. This is equally true in political science, where
comparison is seen as essential. (As in the old joke: ‘Who first discovered
water?’ ‘I don’t know, but it wasn’t a fish.’) Arguably we cannot even
think without comparison and distinction; certainly we cannot think
clearly. Much political theory is thus about making distinctions and
showing how they affect our conclusions – for example, that the posi-
tive/negative liberty distinction excludes important types of liberty, or
that some criticisms of ‘liberalism’ do not weaken all types of liberalism.
Distinctions, unsurprisingly, play a key part in this book. At the most
general level, some chapters will introduce entirely new ideas. I suspect
that for many readers this will include Frazer’s chapter on moral senti-
mentalism and Hamlin’s on positive political theory: simply being
exposed to these ideas, seeing a distinction that one had not been aware
of, will attract some readers to those methods. More specific distinctions
include Quong’s typology of three forms of contractualism, Olsthoorn’s
account of different types of conceptual analysis and Kogelmann and
Gaus’s distinction between the use of rational choice theory for explana-
tory, predictive or normative ends, and in parametric or strategic situa-
tions. But every chapter does something like this: it is what political
theorists and philosophers do.
However, distinctions can constrain. Thus Ackerly and Bajpai
enjoin readers not to see their four types of comparative political thought
as alternatives, while I criticize the distinctions that dominate current
methodological thinking about textual interpretation. Of course, many
of this book’s distinctions will be superseded in time.
14 Chapter 1

Political theorists may need to engage with empirical research,

as argued by Jubb, Schmidtz, Ackerly and Bajpai and Leader Maynard.
Meanwhile, we may already be using some methods without realiz-
ing it, so we should learn to use them well, as argued by Brownlee and
Stemplowska, Frazer, Hamlin and me, and implicitly by Knight and by
Olsthoorn. This is one reason why we should have a broad methods
training in political theory. Another reason is that it may suggest that
we might not be using a method or approach that would in fact
help us. Brownlee and Stemplowska’s argument that good thought
experiments are equivalent to valid philosophical arguments should
make some sceptics more open to thought experiments. Kogelmann
and Gaus encourage political theorists to engage with rational choice
theory. Ackerly and Bajpai want normative theorists to broaden their
horizons by engaging with comparative political thought. Leader
Maynard encourages normative theorists to conduct ideological analysis
on their own ideas. Goodin’s enjoinder to engage with continental poli-
tical theory is something this book does not, alas, deliver on (Ackerly and
Bajpai’s chapter being a partial exception). My chapter suggests that
historians often shut themselves off from philosophical readings of texts.
Frazer reminds us that the rhetorical norms of political theory are recent
inventions: perhaps we have wrongly avoided sentimentalist approaches.
Indeed, every chapter, explicitly or implicitly, justifies its method – and
every chapter, explicitly or implicitly, shows that we may not always be
using current approaches as well as we could. That is of course this
book’s main message.
Research should ideally be problem-driven, emphasize Ackerly
and Bajpai. We should pick methods that best suit the problem at hand
rather than picking problems that ‘our’ method can answer. But in
practice, some scholars simply do see themselves as exponents of
a particular method or approach (e.g. rational choice, textual interpreta-
tion, ideological analysis) and rarely have difficulty finding problems.
Indeed, collective progress might even be greater if problem-driven scho-
lars build on the insights of method-driven experts, just as social scientists
benefit from having quantitative social scientists wedded to a particular
statistical technique or ethnographers skilled at collecting detailed
anthropological data.
This means that the ideal type of research is not ideal for
everyone and would not be ideal if everyone did it. Academia
works best when we study what we love and when there is rigorous
and robust discussion of whether there are gaps or under-emphases in
our collective focus. People might then change their approach or
newcomers can correct these lapses.
Introduction: A ‘How-to’ Approach 15

Contrast this view with John Dunn’s ‘contempt’ for political scien-
tists who do empirical research without considering norms: ‘an adult
and civically responsible human being’ would not study politics like
this (1996: 30–1). Dunn is right that we need more integration of
empirical and normative research. But we do not need or want
everyone to do this: normative researchers may need to draw on
empirical research involving meticulous case studies or exhaustive
statistical number-crunching conducted by scholars with no taste for
normative analysis and no skill in doing it. And those narrow empiri-
cal political scientists may in turn depend on even narrower metho-
dological research by political scientists who are not even interested
in empirical research but merely the methodology of case studies or
of statistical analysis.
So, when Skinner is criticized for overlooking social and religious
contexts, or when Rawls is criticized for overlooking institutional details,
this might merely be a criticism of such oversights, not a recommendation
that Skinner and Rawls should actually have done such research. Let
Skinner be Skinner: why make him start new research projects in social
or religious history when authorities in social and religious history can do
that themselves, building on Skinner’s expertise, filling in his gaps, and
seeing how his conclusions would differ? Let Rawls be Rawls: why make
one of the world’s great political philosophers learn political science when
we can add that to his philosophizing while he concentrates on doing what
he does best? Of course it could be that Skinner or Rawls could spend
a few weeks reading up on other people’s research, and that this would
change their conclusions; but we should not forget that the individual and
the collective ideal are not always identical.
So, we probably want a diversity of approaches – methods-driven and
problem-driven research, specialist and pluralist research. We want peo-
ple who see deeply and people who see widely. So, political theorists
should play to their strengths while recognizing the inevitable limita-
tions of any approach, narrow or wide, and while being aware that
political theory is a collective enterprise where we all stand on each other’s
shoulders, so to speak.
I want to finish with some brief comments on how methodology of
political theory benefits you, and most importantly, with some meta-
reflections on how to do methodology of political theory. There are two
main benefits of thinking about political theory methods. First, good
methods can improve your arguments – the central aim of this
book. Second, and less obvious, methodological reflection can give
you extra publications, e.g. a journal article early in a PhD, or a spin-off
article from a substantive project. This helps your academic career, gives
16 Chapter 1

you a wider audience than you might otherwise get and lets you poten-
tially make a significant contribution to the discipline.
But be careful of writing highly abstract methodological analyses. If this
book convinces you of anything, it is hopefully that methodological
publications should ideally be practical and well exemplified, by
showing what practical guidelines follow from one’s methodological
principles and by illustrating how these ideas have been or should be
implemented. Jubb’s chapter is significant partly because realist political
theory is often very negative: its practical implications have been hard to
grasp. My own chapter reflects my feeling that little methodological
discussion in the history of political thought has actually helped me to
interpret ambiguous passages in texts or to resolve apparent contradic-
tions between different parts of the same text. Indeed, every chapter has
sought to show the value of a practical and well-exemplified approach to
methods in political theory.
This suggests that studying methodology should also be practical
and well exemplified. If using this book to teach methods, one could for
example set students a chapter to read and a substantive application (good
or bad) of the relevant method. This encourages students to think
through the methodological issues and apply them to a substantive issue
or study – which can in turn help us think through the methodological
issues.
So, methodological insights often come from substantive research.
Although methodological developments can happen at a very theoretical
level, or by borrowing from methods in other fields, often one simply reflects
on actual research in political theory and makes inferences about what is
done well or less well. We hope that by reflecting on our methodological
recommendations, and by reflecting on actual research in political theory,
you can contribute to the continuing development of methods in
analytical political theory. That is the spirit in which we have written
this book.

Acknowledgements
For comments and criticisms on earlier versions of this chapter, I thank
Kimberley Brownlee, Alan Hamlin, and my anonymous Cambridge
University Press reviewers.

References
Bentham, Jeremy, 1843. The Works of Jeremy Bentham, volume III, ed.
John Bowring. Edinburgh: William Tait.
Introduction: A ‘How-to’ Approach 17

Bentham, Jeremy, 1996. An Introduction to the Principles of Morals and Legislation,

ed. J. H. Burns and H. L. A. Hart. Oxford: Clarendon Press.
Burke, Edmund, 1999. Select Works of Edmund Burke: Volume 1. Indianapolis, IN:
Liberty Fund.
Cranston, Maurice, 1997. The Solitary Self: Jean-Jacques Rousseau in Exile and
Adversity. University of Chicago Press.
Darwall, Stephen, 2003. ‘Introduction’, in Stephen Darwall, ed., Contractarianism,
Contractualism. Oxford: Blackwell, 1–8.
Dunn, John, 1996. The History of Political Theory and Other Essays. Cambridge
University Press.
Hobbes, Thomas, 1990. Behemoth or The Long Parliament, ed. Ferdinand Tönnies.
University of Chicago Press.
Hobbes, Thomas, 1991. Leviathan, ed. Richard Tuck. Cambridge University
Press.
Hobbes, Thomas, 1994. The Elements of Law, in Human Nature and De Corpore
Politico, ed. J. C. A. Gaskin. Oxford University Press.
Hutchinson, Katrina, and Fiona Jenkins, eds., 2013. Women in Philosophy: What
Needs to Change? Oxford University Press.
Leopold, David, and Marc Stears, eds., 2008. Political Theory: Methods and
Approaches. Oxford University Press.
Machiavelli, Niccolò, 1988. The Prince, ed. Quentin Skinner and Russell Price.
Cambridge University Press.
Mill, J. S., 1974. Essays on Ethics, Religion and Society, ed. J. M. Robson. University
of Toronto Press.
Mill, J. S., 1989. On Liberty and Other Writings, ed. Stefan Collini. Cambridge
University Press.
Miller, David, and Richard Dagger, 2003. ‘Utilitarianism and beyond: contem-
porary analytical political theory’, in Terence Ball and Richard Bellamy, eds.,
The Cambridge History of Twentieth-Century Political Thought. Cambridge
University Press, 446–69.
Reeve, C. D. C., 1988. Philosopher-Kings: The Argument of Plato’s Republic.
Princeton, NJ: Princeton University Press.
Rousseau, Jean-Jacques, 1979. Emile or On Education, ed. Allan Bloom. New York:
Basic Books.
Rousseau, Jean-Jacques, 1997. The Social Contract and Other Later Political
Writings, ed. Victor Gourevitch. Cambridge University Press.
Skinner, Quentin, 2002. Visions of Politics. Volume I: Regarding Method. Cambridge
University Press.
Wolff, Jonathan, 2013. ‘Analytic political philosophy’, in Michael Beaney, ed.,
The Oxford Handbook of the History of Analytic Philosophy. Oxford University
Press, 795–822.
2 How to Write Analytical Political Theory

Robert E. Goodin

1) Show your work. When you are thinking through a problem, you
naturally start by thinking up some plausible line, then considering things
that might be wrong with that, and often rejecting what was superficially
a promising line. Instead of just erasing all of that as a ‘mistake’ or ‘false
start’, write it all out (the original idea, why it was tempting, why it is
wrong). That’s what counts as argument in analytical philosophy.
2) Clean kill/clear clutter. Once you’ve identified your main
argument, clear away superfluous clutter that doesn’t speak to that
argument directly. KISS (‘Keep It Simple, Stupid.’). Remember, ‘clear’
is the highest term of praise in analytical philosophy.
3) Organization = argument. Use the structure of the paper to show
the structure of the argument, wherever possible. If you’ve got several
separate arguments, or complex sub-arguments, put them into new (sub)
sections whenever you can. Let the structure of section breaks do a lot of
work for you.
4) Layer. Once you have a nice clear main line of argument, feel free to
add layers of texture within subsections.
5) Revise, revise, revise. Leave it in the bottom drawer for a month
(maybe while it’s out for comment) and come back to it and revise again.
It’s surprising what you’ll see upon revisiting it with a fresh eye.
6) Let weight on the page be proportional to the weight the
proposition bears in your argument. If something is important,
signal clearly that fact (‘And here’s the crucial point: . . . ’). If it’s
important, go on about it at appropriate length, even at the risk of
redundancy and just rephrasing the same point in other words.
7) Distinctions = arguments. Simply pointing out a distinction –
drawing attention to the fact that this is different from that (with which
it might naively be confused or conflated) – is the most typical way of
making arguments in analytic philosophy.
8) Introduce key devices early. Where there is some important dis-
tinction or device that you’re going to need later, particularly at multiple
points in the paper, it can be a good idea to introduce it early. If it’s
important, and you want people to notice it, don’t just sneak up on it

18
How to Write Analytical Political Theory 19

without warning, and leave some important point buried somewhere in

the middle of the paper where it might get overlooked.
9) Intuition pumps require care. The grandest ambition of analytic
philosophy is to get support for some proposition about which we feel less
confident by showing that it follows from some other proposition in which
we’re more certain. (Think of Descartes, tracing everything ultimately
back to the proposition of which he felt most certain, ‘Cogito ergo sum.’)
The way that in practice plays out in applied moral philosophy is via
‘intuition pumps’, examples (often contrived, artificial) about which we
have a strong intuitive response; and then assimilate other cases of interest
to those. The strategy often misfires, for various reasons. One is that the
cases described are ‘crazy’, atypical of real life and not ones around which
our intuitions have been honed – so our intuitive responses to them are
not to be trusted. Another classic way that strategy misfires is that the
analogy to real-world cases is flawed in some way.
10) Horses for courses. Analytic philosophy is not the only mode.
There are clear cases in which, and audiences for which, you should
write in a more lugubrious and less clipped mode. Sometimes pure
random free association is a good way of generating ideas that, stepping
back from which, you can then clean up and put into proper analytical
order.1 Letting continental philosophy wash over you can sometimes be
useful, in just that way.
11) Getting started. A good way to start a new project is to compile
a whole jumble of dot-points and short paragraphs containing random
ideas that you might include in a paper. Then look across them all to see if
there’s some ‘story’ you can draw out of those (drawing out some of them,
throwing away others, typically).
12) Ask around. Both before writing and after, don’t hesitate to send
things to colleagues who might give you good background information,
advice or comments. But do exercise quality control: people who are not
very good give you advice that’s not very good. (Sometimes you have to
circulate papers to such people out of courtesy, of course: just remember,
‘not all comments are created equal’.)
13) Feel free to ignore advice, even good advice. Don’t overreact to
advice. You can make a paper a lot worse by revising it too much – by
trying to take on board all advice you get on it. This obviously might
happen when you try to be too accommodating, taking on board points
that don’t really fit just to be nice to someone who has gone to the trouble
to comment. But it can happen too when you revise to take into account
points that you agree are genuinely good and important points – even so,

1
This is what James G. March (1976: 76–81) celebrates as ‘the technology of foolishness’.
20 Chapter 2

the revisions can sometimes pull the paper all out of shape. If they’re really
important, you might want to rewrite the whole paper (or whole section of
the paper) afresh. But often retroﬁtting them to the existing structure just
doesn’t work.
14) Placement. Every paper has a ‘natural home’. It’s just a case of
matchmaking, knowing the personality of a journal and whether this
paper would be a good ﬁt.

Acknowledgements
These notes grew out of discussions with Chiara Lepora during our time
together at the National Institutes of Health, 2009–10.

Reference
March, James, 1976. ‘The technology of foolishness’, in James March and
Johan Olsen, eds., Ambiguity and Choice in Organizations. Bergen:
Universitetsforlaget, 69–81.
3 Thought Experiments

Kimberley Brownlee and Zoﬁa Stemplowska

1 Introduction
A thought experiment is, in one sense, just what its name suggests – an
experiment in thinking. But it is thinking of a distinctive, imaginative
kind that offers a potentially powerful investigative and analytic tool in
mathematics, science, and philosophy. In science, thought experiments
are a well-accepted, uncontroversial mechanism for testing hypotheses,
and in mathematics, they are one of the principal tools for valid reason-
ing. They can build and destroy arguments (Brown 1986). In negative
terms, they can (1) expose a contradiction, (2) undermine a key pre-
mise, (3) reveal a conflation of concepts or principles, or (4) highlight
the counterintuitive implications of an argument. In positive terms, they
can (1) demonstrate the consistency or coherence of a set of principles/
concepts, (2) highlight congruities and similarities between different
claims, (3) reveal the scope of the application of a given principle, and
(4) bring forth intuitions not previously considered, amongst other
things.
In philosophy, some thought experiments are highly influential, even
famous. In moral and political theory, famous examples include the
following:
1. Philippa Foot’s/Judith Jarvis Thomson’s Trolley Problem:
A runaway trolley is hurtling towards five people who are working on
the railroad track up ahead. The driver can either continue onto the
track ahead, thereby killing the five, or steer onto a second track off to
the side on which only one man is working, thereby killing the one. Is it
permissible to turn the trolley? (Foot 2002: 23; Thomson 1985:
1395).
This thought experiment serves various purposes. Some argue, for exam-
ple, that it is permissible to turn the trolley because the negative duty not
to kill the five outweighs the negative duty not to kill the one. This
experiment is discussed in a variety of political theory contexts, including
in just war theory on the legitimacy of defensive harm.

21
22 Chapter 3

2. Philippa Foot’s/Judith Jarvis Thomson’s Trolley Problem and

Transplant Case: A doctor can save the lives of five dying patients
by killing one healthy person and giving her organs to the five. If it is
permissible to turn the trolley, is it also permissible to kill the one to
save the five? (Foot 2002: 24; Thomson 1985: 1396).
This thought experiment also serves various purposes. Some thinkers
argue, for example, that unlike the trolley problem, it is impermissible
to kill the one healthy person in the transplant case because the positive
duty to save the five outweighs the negative duty not to kill the one.
3. G. A. Cohen’s Camping Trip: Suppose that a group of friends goes
camping together. The campers could either contribute according to
their abilities and resources, with the expert fisher going fishing and
the skilled forager finding apples, and so on. Alternatively, they could
each assert their rights over their own equipment and talents and use
them to bargain with the other members of the camping group. Is it
better to base a camping trip on the principles of market exchange and
private ownership or on the socialist principles of collective ownership
and planned mutual giving? (Cohen 2009: 3ff.).
Cohen uses this thought experiment as his starting point to challenge the
idea that socialism is infeasible and counterintuitive, and to consider
whether societies are relevantly distinct from camping trips, or whether
societies that cultivate the mechanisms to harness human generosity
could be governed by socialist principles.
4. John Rawls’s Original Position: We are asked to put ourselves
in the position of free and equal persons collectively deciding on
and committing to a set of principles of justice for society.
To ensure our impartiality as deliberators, we engage in this
process behind a ‘veil of ignorance’ that shields us from the
knowledge about who we will be in the society and what advan-
tages and disadvantages we will have. What principles would we
choose? (Rawls 1971).
This thought experiment is used to offer support for a liberal-egalitarian
conception of justice.
5. Robert Nozick’s Experience Machine: Suppose that you could put
yourself into an experience machine for the rest of your life, which
would give you all the experiences you find enjoyable and valuable
without your knowing, once you were in the machine, that these
experiences were not real. Would entering the machine be a good
choice from the point of view of well-being? Would it give you all
that mattered? (Nozick 1974: 42–5).
This thought experiment is used to challenge the claim that the realm of
value is exhausted by hedonistic pleasure. Nozick argues that the realm of
Thought Experiments 23

value and well-being is not exhausted by the pleasure of ‘experiences’; we

care about whether our experiences are real.
6. Robert Nozick’s Utility Monster: Suppose there is a person who gets
enormously greater gains in utility from any sacrifice of others than
those others lose in utility through their own sacrifice. Is it morally
required or morally permissible to sacrifice these persons for the sake
of the monster’s greater utility? (Nozick 1974: 41).
This thought experiment is used to challenge the intuitiveness of hedo-
nistic utilitarianism.
7. Ticking Bomb: Suppose that you are an investigator who has
a suspect in custody who you know has planted a bomb somewhere
in your city. The bomb will soon explode. Suppose also that torturing
this person will give you the information you need to locate and
neutralize the bomb. Is it either morally permissible or required to
torture the person in custody? (adapted from Walzer 1973: 166–7).
This thought experiment is used to gauge the permissibility of torture.
(See the Appendix for other examples of thought experiments.)
Unlike in mathematics and science, in political theory and in philoso-
phy in general, the use of thought experiments is a matter of lively
controversy. Two especially pressing objections against their use are the
following:
1. The Objection of Bias: Thought experiments both invite systematic
bias and entrench existing biases.
2. The Objection of Inherent Ambiguity: Thought experiments often are
inherently ambiguous, leading to inescapably opaque judgements.
These objections are troubling because they challenge the very possibility
of making logically and philosophically respectable use of thought experi-
ments. Neither objection is forceful in its general form because, if it were,
it would impugn the less controversial use of thought experiments in
mathematics and science and not just philosophy. These two objections,
however, may be thought to target the use of thought experiments in sub-
disciplines of philosophy such as political theory and moral theory where
thought experiments are deployed not only for conceptual and logical
purposes, but also for normative and evaluative purposes. Using thought
experiments in political and moral theory in particular may seem suspect
because such thought experiments abstract away from and idealize real-
life cases or even invent fantastical scenarios, but nonetheless purport to
guide real-life behaviour.
The use of thought experiments in political and moral theory is also
subject to further, less weighty objections. First, such thought experi-
ments are said to be in poor taste since they often involve fantastic
scenarios of suffering, death, and cruelty that trivialize that suffering.
24 Chapter 3

(Ironically, thought experiments of the past were also criticized as too

trivial: see Winch 1965: 199–200.) Second, they are said to impoverish
our understanding of urgent problems, as they are devoid of rich social
context (O’Neill 1986: 12, 20–1; but note that O’Neill herself does not
attribute this feature to thought experiments in general, only in some
contexts within political and moral theory). Third, thought experiments,
such as the Ticking Bomb, are said to misrepresent the vast majority of
relevant real-life cases and thus create the false impression that the world
is simpler and more manageable than it is. These latter three objections
can be set aside, however, because their force, while somewhat doubtful,
could be granted without abandoning the practice of thought experiments
in political and moral theory. They seem to invite theorists to engage in
careful and tactful delineation of thought experiments rather than to
abandon them altogether.
The main purpose of this chapter is to provide a guide for the use of
thought experiments in political theory (although what we say holds more
generally for normative theory). As part of that objective, we aim to refute
the Objection of Bias and the Objection of Inherent Ambiguity against
thought experiments in this area. A further, related purpose of this chap-
ter is to flesh out positively the distinctive argumentative value that
thought experiments have in normative theory. We begin by distinguish-
ing the concept of a thought experiment from things with which it is some-
times conflated, namely, introspective psychological experiments and other
argumentative tools that appeal to the workings of the imagination such as
descriptive hypothetical examples (Section 2). We then respond to the
Objection of Bias and the Objection of Inherent Ambiguity, first by
articulating and defending a set of necessary, formal conditions for for-
mulating well-posed thought experiments in normative theory
(Section 3), and second by showing that these conditions do not preclude
the use of thought experiments that involve practical impossibilities or
imaginatively opaque components (Section 4). We conclude by high-
lighting the key ‘how-to’ instructions for designing thought experiments
in political and moral theory (Section 5).

2 Deﬁnitions
Thought experiments have been characterized variously as devices ‘of the
imagination used to investigate the nature of things’ (Brown and Fehige
2014), picturesque arguments (Norton 1996: 334), purely mental proce-
dures that aim to reveal something about the relationship between two or
more variables (Sorensen 1992: 186, 205), and judgements about what
would be the case if the particular state of affairs described in some
Thought Experiments 25

imaginary scenario were actual (Gendler 1998: 398). Our conception of

thought experiments in normative theory is as follows:
A thought experiment is a multi-step process that involves (1) the mental visualiza-
tion of some speciﬁc scenario for the purpose of (2) answering a further, more
general, and at least partly mental-state-independent question about reality.1
The reference here to ‘mental visualization’ highlights the imaginative
quality of thought experiments. They are not purely abstract or formal
operations of thought. Rather, they are operations of thought structured
to invite visualization. This does not mean that, by their nature, thought
experiments cannot intelligibly and proﬁtably deploy concepts that defy
visualization, such as a square circle, a world with different laws of nature,
or an episode of giving birth to oneself. Rather, the point in highlighting
the visual quality of thought experiments is to note that they are not
carried out purely at the level of abstract principle, but instead ‘invoke
particulars irrelevant to the generality of the conclusion’ (Norton
1996: 336) to be drawn from their use.

2.1 Descriptive Hypothetical Examples versus Thought Experiments

The reference to ‘mental visualization’ should not obscure the fact that
thought experiments are only a subset of a broader category of hypothetical
scenarios that involve visualization and imagination. A second subset of
that category is descriptive hypothetical examples, which, unlike thought
experiments, neither test nor contribute an independent step to a chain
of reasoning. Purely descriptive hypothetical examples, such as ‘I have in
mind here someone like Anna Karenina,’ or ‘God is an example of a perfect
being,’ or ‘Annette is a person who is so poor that her cupboard is bare,’ are
elucidatory not argumentative. Descriptive hypotheticals and thought
experiments have different functions. The former set the parameters of
the type of problem under consideration and/or clarify the concepts at
issue. The latter either are independent argumentative moves’ or they
test, and hence support or undermine, argumentative moves.
Although we shall not examine descriptive hypothetical examples in
what follows, it is worth noting two features of them in relation to
thought experiments. First, descriptive hypotheticals can be proto-
thought experiments, that is, sometimes they can be easily developed

1
We do not mean to settle the debate between expressivists/non-cognitivists on the one
hand and cognitivists on the other. Even if normative judgements are ultimately entirely
a matter of affective states (and hence are not mental-state-independent), we mean to
signal that thought experiments aim to provide answers that at least appear to be partly
mental-state-independent.
26 Chapter 3

into thought experiments. For instance, once we begin to describe

Annette’s situation to specify the type of poverty we wish to examine,
we can also use that description to test the acceptability of various
responses to her plight. Hence, we might ask: ‘Would we be prepared
to leave someone so impoverished to struggle on her own?’ Our initial
description of Annette’s impoverishment is not a thought experiment,
but it opens up the prospect of posing questions about how to treat
Annette.
Second, unlike thought experiments, descriptive hypotheticals can
assume what they are meant to illustrate. This would be a fatal problem
for a thought experiment as part of a chain of reasoning, but not for
a descriptive hypothetical as an elucidatory device. We return to this in
Section 3.1.1.

2.2 Psychological Experiments versus Thought Experiments

The second part of our conception of a thought experiment – that its function
is to answer a further, more general, and at least partly mental-state-
independent question about reality – allows us to distinguish thought
experiments from introspective psychological experiments (or ‘psychological
experiments’ for short) (Unger 1982: 117, Sorensen 1992: 67, Brown
2014). The latter are mental procedures that aim simply to predict or to
reveal to us our psychological/mental states. A psychological experiment
asks such things as: ‘Can you make yourself believe that you are an oyster?’,
‘Can we imagine what it is like to be a bat?’, ‘Putting aside whether it is
permissible, would we actually be able to bring ourselves to turn the
trolley?’, or ‘How would you feel if your child were killed?’ Psychological
experiments are a distinctive kind of mental experiment in which the
generation of a given mental state is precisely and uniquely what is being
tested. For instance, when you ask someone whether, in circumstance C,
she would fear an attacker enough to kill him, your aim is to ascertain
through this test what her mental state is likely to be in such circumstances
(or at least what she thinks it is likely to be). By contrast, when you ask an
accountant what is 1,236 divided by 3, ascertaining her mental state is not
normally the object of the ‘experiment’ (unless you wish to ﬁnd out how she
will react). The object of the experiment is to get at some feature of the
world – the answer 412 – that is independent of her mental state.
The commonly asserted claim that thought experiments, such as the
ones outlined above, generate strong intuitions invites confusion between
thought experiments and psychological experiments because it can be
read to imply that all that thought experiments are meant to test are
affective (psychological) states. But the confusion between thought
Thought Experiments 27

experiments and psychological experiments may also have a deeper

source in that some thought experiments necessarily include psychologi-
cal experiments as a preliminary step in order to reach further conclu-
sions. This occurs when (and because) the variables that a given thought
experiment examines include or depend on psychological states, usually
ones involving emotions (i.e. affective states). For example, take the
following thought experiment:
Attacker: Suppose that we see a person being attacked. And suppose that we are
morally required to call the police when we see a person being lethally attacked.
If the police cannot arrive in time, are we also morally required ourselves to kill the
attacker (assuming that our action will not threaten the institution of policing)?
In order to engage with this thought experiment, it may be necessary
amongst other things to run a psychological experiment by asking our-
selves if we would be able to bring ourselves to kill the attacker. (Would
you?) We might want to ask this question if, say, we accept that ‘ought
implies psychological can’; i.e., if we were psychologically unable to bring
ourselves to kill the attacker, then, if ought implies psychological can, we
would not be morally required to do so. Nonetheless, although this
psychological experiment is part of Attacker, that thought experiment is
not exhausted by the performance of the psychological experiment since
the thought experiment requires us in addition to reach a judgement about
a moral requirement. That is, it requires us to reach a judgement about
what is morally required of us in this kind of case (and that judgement is,
on standard objective conceptions of morality, at least partly independent
of our beliefs as the agent).
Given that psychological experiments in political and moral theory
usually test affective states, one rough and ready way to distinguish thought
experiments from psychological experiments in this area is to think of
thought experiments as answering ‘What is your moral judgement?’ and
psychological experiments as answering ‘How would you feel?’2 We stress
the distinction between thought experiments and psychological experi-
ments for several reasons. First, it allows us to explain the nature of the
intuitions that thought experiments are designed to elicit. Simply put,
unlike most psychological experiments, thought experiments are not
intended to elicit raw, unreﬂective intuitions or brute reactions. Their
results can and often should be the fruit of reﬂection (see Knight’s chapter

2
One might worry that an expressivist or non-cognitivist would reject this as a false
distinction. But a sufﬁciently sophisticated version of non-cognitivism presumably accepts
that, even if moral judgement is ultimately a matter of affective states, there is nonetheless
a plausible distinction to be drawn between raw affective states and ‘gardened’ or reﬂective
ones.
28 Chapter 3

on reﬂective equilibrium in this volume). Second, it has implications for the

way in which data gathered through thought experiments can enter into
normative reasoning. If what matters in thought experiments are not (or
not exclusively) raw affective states, then there is more room for rational
debate over the appropriate response to a given thought experiment.

2.3 Simple Thought Experiments versus Complex Thought Experiments

Within the category of thought experiments, there are some important
distinctions to be drawn. The first is between simple and complex thought
experiments. A simple thought experiment, such as the Trolley Problem,
considers a single scenario. In political and moral theory, simple thought
experiments tend to raise questions of whether some action is morally
wrong, permissible, or obligatory, or whether some state of affairs is fair,
equal, just, good, and so on. For instance, the Trolley Problem raises the
question of whether it is permissible to turn the trolley and divert the harm
from the five to the one. Oftentimes, the theorist’s intuitive, though not
unreflective response to such thought experiments is taken to be evidence
for or against the hypothesis being tested in the thought experiment (e.g.
that turning the trolley is morally permissible/required).
By contrast, a complex thought experiment, such as the combination of
the Trolley Problem and Transplant Case, considers two or more scenar-
ios in relation to each other.3 This complex experiment – Trolley Problem
and Transplant Case – contrasts (the simple thought experiment) Trolley
Problem with (the simple thought experiment) Transplant Case. It aims
to establish whether our normative answers in the one case align with our
answers in the other case. In political theory, complex thought experi-
ments serve various negative and positive argumentative purposes, such
as exposing a disanalogy or confirming an analogy, undermining or
affirming a hypothesis, revealing a conflation of concepts or principles,
and bringing to light unacknowledged intuitions.
This distinction between simple and complex thought experiments is
significant because some simple thought experiments need not satisfy the
condition of validity (see Section 3.1.2) that applies to all complex
thought experiments.

2.4 Contingency, Necessity, and Imaginability

The ﬁnal distinctions to highlight within the category of thought experi-
ments relate to their degree of practical possibility and of imaginative
clarity.

3
Paradigmatically, complex thought experiments involve pairwise comparisons.
Thought Experiments 29

Thought experiments can be more or less practically possible.

The category of hypothetical is a continuum that includes both the likely
and probable though non-actual at one end, and the extremely unlikely
and even the impossible at the other end. The former can be described as
contingently hypothetical (e.g. ‘Imagine that you are walking by a pond and
spot a drowning child’). Such hypotheticals can be structured around
actual events, such as rescue boat operators in New Orleans after
Hurricane Katrina having to choose between rescuing the five adults on
one rooftop or the three children on another, but they remain hypothe-
tical in that they abstract away from the real case. One reason to construct
contingently hypothetical thought experiments even when real-world
cases are readily available is that often the mention of a real-world exam-
ple invites protracted discussion of the facts, which can distract from the
problem at hand.
At the other end of the continuum lie thought experiments that can be
described as necessarily hypothetical (e.g. ‘Imagine a spear flying towards
the edge of the universe’) or at least necessarily hypothetical for us here and
now (e.g. ‘Imagine a society that has eliminated poverty’). Thought
experiments that fall closer to this latter end of the spectrum are con-
troversial to some because they depart significantly from our lived, every-
day reality. Being necessarily hypothetical in either of the two senses just
noted is one way in which a normative-theory thought experiment may be
said to be ‘wacky’.
Another way in which a normative-theory thought experiment may
be ‘wacky’ is in being imaginatively opaque. Robert Nozick’s Utility
Monster involves an imaginatively opaque being since his, her, or its
pleasure in sacrificing others must be of a fantastic quality to outweigh
the acute suffering of all of those who are sacrificed. As ordinary
creatures with ordinary abilities for happiness, we are unable to ima-
gine the kind of fantastic happiness that such a being would have to
feel in order to outweigh the suffering it caused to those it sacrificed
(Parfit 1986: 389).
These two senses (and sources) of wackiness – necessary hypotheti-
cality and imaginative opacity – can overlap but are conceptually
distinct, since some cases of practical impossibility, such as my jump-
ing 100 feet in the air, are nonetheless readily imaginable, while some
cases of imaginative opacity, such as the experiences of sleepwalking
and of insanity, are nonetheless readily practically possible and indeed
actual. Commonly cited examples of wacky thought experiments
include Rawls’s Original Position, Nozick’s Experience Machine,
and Nozick’s Utility Monster. Since imaginatively opaque and neces-
sarily hypothetical thought experiments invite the most controversy in
30 Chapter 3

normative theory, they will be the main focus of our defence of

thought experiments in Section 4.

3 Necessary Conditions for Well-Posed Thought

Experiments
What would a well-posed thought experiment in political theory
look like? In this section, we outline and defend two necessary
conditions for well-posed thought experiments in normative theory:
philosophical respectability (Section 3.1) and argumentative rele-
vance (Section 3.2). In broad terms, these conditions of well-posed
thought experiments apply outside of normative theory, but they are
particularly salient to normative theory. Although both conditions
apply to both simple and complex thought experiments, the ﬁrst
condition places different constraints on each type of thought
experiment.
These conditions are a non-exhaustive set in the sense that there are
further conditions that any good argument must meet (e.g. clarity), which
we do not mention, as we wish to focus on what is special to thought
experiments in particular. We believe that thought experiments that
satisfy these conditions will be genuinely well-posed provided they do
not fall foul of conditions that apply more generally to philosophical
investigation.

3.1 Philosophical Respectability

This condition has two distinct dimensions, the ﬁrst of which is non-
question-beggingness (Section 3.1.1), which applies straightforwardly to
both simple and complex thought experiments. The second dimension,
validity, applies to all complex thought experiments and to some simple
ones (Section 3.1.2).

3.1.1 Non-Question-Beggingness Thought experiments should not

assume an answer to the question they pose. So, when formulating
thought experiments, one cannot assume that it is permissible to
torture the bomb planter in order to argue that it is permissible to
torture him.
Of course, it is not always obvious when answers are being assumed and
questions are being begged. Consider Wittgenstein on the ontology of
moral judgements:
Suppose one of you were an omniscient person and therefore knew all the move-
ments of all the bodies in the world dead or alive and that he also knew all the
Thought Experiments 31

states of mind of all human beings that ever lived, and suppose this man wrote all
he knew in a big book, then this book would contain the whole description of the
world; and what I want to say is, that this book would contain nothing that we
would call an ethical judgment or anything that would logically imply such
a judgment. (Wittgenstein 1965: 6)
Wittgenstein’s scenario presupposes that there are no facts in the
world of the kind that he aims to deny and hence his hypothetical
does not test the claim that this is the case. Any other conclusion
about the book’s content would be contradictory. This would be
a fatal problem for a thought experiment testing whether ethical
judgements are facts. However, the example would be acceptable as
a descriptive hypothetical that merely aims to elucidate what
Wittgenstein means by facts.

3.1.2 Validity When we ﬁrst pose a thought experiment to ourselves,

we should pose it as an open question, in the way that all of the thought
experiments presented thus far have been posed. However, the question
hopefully gives rise to answers, that is, the results of the thought experi-
ment. Results are broadly of two types. First, they may simply consist in
answers about what is morally required, permissible, etc. (e.g. it is
impermissible to kill one to save ﬁve in the Transplant Case). Second,
they may consist in such answers together with a further hypothesis
about why this is the correct answer (e.g. because harming is worse
than not aiding). Simple thought experiments allow but do not require
the researcher to propose the hypothesis. All complex thought experi-
ments, however, necessarily contain at least an implicit hypothesis about
what does the work in one of the simple thought experiments; the next
simple thought experiment is then added precisely in order to test that
hypothesis (see later in this chapter for more on this).
Where a thought experiment contains or generates a hypothesis
explaining our intuitive reactions to the scenarios involved, the
thought experiment can be ‘translated’ into an argument (See also
Norton 1996). That argument must satisfy the condition of validity.
In other words, the argument should not involve logically fallacious
reasoning. Of course, what constitutes fallacious reasoning is
a matter of some debate. The point is simply that thought experi-
ments that contain and generate hypotheses can and should be held
to the same standards of valid reasoning as conventional arguments,
whatever those standards may be. Inability to translate such thought
experiments into a valid argument would indicate that we are unsure
either of what the experiment is supposed to test or of whether it
presupposes what it is meant to reveal.
32 Chapter 3

An example of a thought experiment that satisﬁes the validity condition

is the following from Peter Singer.
The Pond and the Envelope: Imagine that you are walking by a shallow and
isolated pond in which you spot a drowning child. You can easily save the child.
Must you save the child? Imagine next that you receive a letter from a charity such
as Oxfam asking you for a donation that you can easily make, to save a child (or,
likely, many children) abroad. If you accepted that you must save the child in the
pond case, must you also save the child(ren) in the envelope case? (Singer 1972:
231–2)
The thought experiment is used to show, amongst other things, that it
makes no moral difference whether the person we can help is
a neighbour’s child ten yards away or a foreigner whose name we shall
never know 10,000 miles away.
Assume that, following Singer, you answer all questions in the
afﬁrmative. Once we form an intuitive afﬁrmative response to the
questions, this thought experiment can be readily translated into
a valid, conventional argument as follows:
p1: We can easily save the child in the pond.
p2 : We have a duty to save the child in the pond.
p3 : The best explanation for P2 is that we have a duty to save others when we can
do so at little cost to ourselves.
c1 ( t h e hy po t h e s i s) : We have a duty to save others when we can do so at
little cost to ourselves.
p4 : We can send money to Oxfam at little cost to ourselves.
p5 : We can save others by sending money to Oxfam.
c2: We have a duty to send money to Oxfam.

The condition of validity is satisﬁed here since the argument into

which the complex thought experiment is translated tracks what the
thought experiment was intending to test or establish and satisfies the
criteria for a valid argument. More generally, of course, Singer would
want us to see the argument not only as valid, but also as sound: he
would want us to accept that there is no relevant difference between
the Pond and the Envelope scenarios in that both require the same
moral response and both are explained by the same general principle
(the hypothesis: C1).
A thought experiment that does not satisfy the condition of validity is
the following:
A Sibling and a Stranger: Imagine that your sibling contracts malaria and
can be saved only if you agree to finance expensive medical treatment invol-
ving a helicopter ride. You can finance it, albeit it will cost you a lot and you
won’t be able to go on holiday for a few years. Must you do it? Imagine next
Thought Experiments 33

that your sibling is healthy, but you can ﬁnance similar life-saving medical
treatment for a stranger. If you accepted that you must save your sibling, must
you not also save the stranger?
This thought experiment can be readily translated into a fallacious
argument:
p1: We can save our sibling at high cost to ourselves.
p2 : We have a duty to save our sibling.
p3 : The best explanation for why we have a duty to save our sibling is because this
is our sibling.
c1 ( t h e h y po t h e s i s ) : We have a duty to save our sibling even when this
involves a high cost to ourselves.
p4 : Saving strangers would involve high costs.
p5 : The costs of saving the strangers would be identical to those of saving our
sibling.
c 2 : We have a duty to save strangers even when this involves a high cost to
ourselves.
This argument is invalid because, even were the premises all true, the
conclusion need not be true: it does not follow from the fact that we have
a duty to save a sibling at high cost to ourselves that we necessarily have
a duty to do other things that are equally costly.
The condition of philosophical respectability has the virtue of
demystifying the status of thought experiments. If thought experi-
ments can be represented as conventional arguments that meet the
standards of valid reasoning, then it is unsurprising that they can act
as solutions to philosophical problems. As we argued, this is the case
with all complex thought experiments and with at least some simple
thought experiments. To be genuinely well-posed, however, thought
experiments should also be analytically useful to us and not corrupt
our reﬂections. This is addressed by the second condition, the con-
dition of argumentative relevance.

3.2 Argumentative Relevance

Thought experiments should be designed in such a way that we can focus
on the relevant aspects of the scenario. We do not want our intuitive
answers to respond to features of the scenario that are not part of the test
and that thereby pollute it. For example, when testing a given hypothesis
(such as a hypothesis about how we ought to treat strangers), it is
necessary not to construct scenarios that more plausibly test an alternative
hypothesis (such as a hypothesis about how we ought to treat our sib-
lings), as we cannot assume that they will elicit the same answers. For
34 Chapter 3

instance, the Sibling and the Stranger could be translated into a valid
argument that fails to meet this condition:
p1: We can save our sibling at high cost to ourselves.
p2 ’: The best explanation for why we have a duty to save our sibling is because
we have a duty to save others even at a high cost to ourselves.
c 1 ’ ( t h e h y po t h e s i s ) : We have a duty to save others even when this involves
high cost to ourselves.
p3 : Saving strangers would involve high costs.
p4 : The costs of saving the strangers would be identical to those of saving the
sibling.
c2: We have a duty to save strangers even at a high cost to ourselves.
The argument is valid, but ridiculous: P2’ misidentifies the principle to be
derived from considering the case of the sibling. Similarly, looking at
a simple thought experiment, if, in the Transplant Case, we forbid the
doctor to kill the one person to save the five on the grounds that a doctor
may never kill, then we are not testing what the experiment is meant to
test, which, among other things, is whether there is a normatively
significant difference between killing and letting die (see likewise Rivera-
Lopez 2005). By prohibiting the doctor from killing, we block the relevant
test, as we allow her status as a doctor to infect our reflection on the
scenario.
All in all, if we want to use our thought experiments as evidence for or
against a given hypothesis (premise), we need to make sure that the results
of the experiment actually support or challenge the argumentative move in
question. The key question is whether it is possible to make this condition
of testing the relevant hypothesis more concrete beyond prohibiting
obvious shifts in focus. We argue that the condition of argumentative
relevance translates into two weak constraints on the design of thought
experiments. The first requires that the experiment allow for rudimentary
alternatives (the Rudimentary Alternatives Constraint). The second
requires that the experiment not encourage narrative-framing bias (the
Moderate Narrative-Framing Constraint). These two weak constraints
can be contrasted with more demanding variants, which we reject.

3.2.1 Rudimentary Alternatives Constraint When we assume the

absence of some (believed) necessary feature of the world, we should
stipulate at least a rudimentary alternative. The aim here is to eliminate
the bias that may come from continuing to assume that the feature still
obtains. For example, an ancient philosopher who believes that objects
can only move by willing to move should not run a thought experiment
like the following:
Thought Experiments 35

Unwilling Rock: Assume that a large rock is not willing to move, but still moves.
Is the rock to blame when it kills someone?
The ancient philosopher who holds that willing is a necessary condition
for moving should not run this thought experiment – without some extra
stipulations – because he is pre-committed to the view that the object that
moved must have been willing to move. He should stipulate instead
a rudimentary alternative for how the object moved; for example, the
object moved because it fell just like a human being might fall if pushed
by a gust of wind.
Unwilling Rock 2: Assume that a large rock is not willing to move, but still
moves, pushed by a gust of wind (against its will). Is the rock to blame when it kills
someone?
Although we endorse the Rudimentary Alternatives Constraint, we
reject the more demanding Fleshed-Out Alternatives Constraint, which
holds that allowed alternatives must be fully fleshed out and rendered
comprehensible to us given what we know about the world (Wilkes
1988: 43ff.). Returning to the ancient philosopher, unlike the
Rudimentary Alternatives Constraint, the Fleshed-Out Alternatives
Constraint would require him to explain how a large, heavy rock can
be pushed by a gust of wind. We acknowledge that a fully fleshed-out
alternative would protect us from certain biases, but the protection is too
restrictive. It is implausible to hold that we really need a fleshed-out,
clear statement of the alternative to the ruled-out feature of the scenario,
in order to prevent the ruled-out feature from determining our
conclusions.
All in all, then, we accept that unconscious bias is real bias. But the
possibility of bias is not a reason to abandon theorizing that might be
subject to it. It is a reason to guard against it within the parameters of the
case. We think that the Rudimentary Alternatives Constraint allows us to
do so.

3.2.2 Moderate Narrative-Framing Constraint We also support, more

tentatively, the Moderate Narrative-Framing Constraint that guards
against thought experiments that encourage narrative-framing bias. For
instance, a thought experiment that draws its scenario from a well-known
novel, event, film, genre, cultural myth, or icon can bring with it con-
siderable narrative baggage in that the context of its creation has its own
purposes that might subordinate or undermine clear reflection on the
scenario as a thought experiment. The problem is best explained with the
use of an example that we owe to Roy Sorensen (1992: 264; see also Parfit
36 Chapter 3

1986: 199–200; Coleman 2000: 58–60). Consider teleportation. Since it

is almost exclusively encountered in the context of sci-fi adventures such
as Star Trek and Harry Potter, its context makes demands of narrative
unity on our reading of teleportation scenarios. As viewers, we want to
believe, for the sake of the story, that it is the same person who is
teleported rather than a new person who is created by the process, and
this may infect the philosophical use we seek to make of teleportation
scenarios.
Sorensen goes on to suggest less plausibly that Nozick’s Experience
Machine also may be systematically distorted for a similar reason,
namely, that we approach this thought experiment as a story about some-
one entering an Experience Machine and we find the possibility of such
a story so unbearably boring that we reject it as a legitimate prospect. But
Sorensen’s position on this is implausible. There is no putative demand of
narrative unity about Experience Machines that necessitates that an
Experience Machine scenario be irretrievably boring. We can rewrite
this kind of scenario as an exciting, Matrix-style adventure that eliminates
the supposed anti-boredom bias. Our rewriting may introduce a pro-
excitement bias in favour of the adventure, but that suggests that we
need only find a middle-of-the-road description of the Experience
Machine experience.
The same is true presumably for most thought experiments. Narrative
bias need not hopelessly infect thought experiment scenarios provided
that we are attentive to the structure of the scenario and to the narrative
assumptions that it can imply. Thus, for example, when we involve the
Nazis to make a point against the permissibility of medical experimenta-
tion, we should be careful not to appeal just to the horrors that the
mention of Nazis invokes.
By endorsing the Moderate Narrative-Framing Constraint in this
form, we reject the more demanding Extreme Narrative-Framing
Constraint that requires thought experiments to be ‘maximally con-
servative’ and lie exclusively within the realm of contingent hypotheti-
cals and never that of necessary hypotheticals, the spectrum of which
we discussed in Section 1 (see Haggqvist 1996: 147; Rivera-Lopez
2005). The central idea behind maximal conservatism is that experi-
ments that require us to depart from standard circumstances that we
would encounter in our world will not track our reactions to the
features of the case as set out in the experiment, but will instead track
our reactions to the standard features of a case encountered in the
actual world. For example, when asked to assess whether to kill one
to save five in the Transplant Case, we will ultimately not be able to
take on board the stipulation that the alternative deaths really are
Thought Experiments 37

certitudes, since in our common experience we may hope that the

five would still have a chance of surviving since we never know for
certain. This alleged limitation of our mental abilities calls into doubt
the usefulness, or even the possibility, of wacky thought experiments
such as necessary hypotheticals.
But, while the problem of polluted intuitions is genuine, the
assumption underlying the postulate of maximal conservatism is
mistaken. The postulate rests on the wrong-footed assumption that
we are likely to reach better judgements when we operate within
a familiar, real-life context than when we operate within an unfami-
liar context. But this is not generally true. We might be better able to
react only to the variables that the thought experiment is meant to
test when the case is set in the context that we do not normally
encounter and are not familiar with, just as a non-native English
speaker might be quicker than a native speaker to spot certain
linguistic patterns in English, or a person unfamiliar with a given
family’s dynamics might be quicker than family members are to spot
instances of mental abuse and exploitation. A radically unfamiliar
context may well make us more attentive to the features of the
scenario that matter precisely because we are less likely to smuggle
in additional assumptions.
That said, we do not deny the potential legitimacy of the worry
about narrative bias. A thought experiment that asks us to assume
a hateful baby or a saintly Mafioso might be hard to execute cor-
rectly. Similarly, we might also worry in relation to the ticking-bomb
thought experiment that it asks us to assume what is very hard to
imagine, namely that the torturer will be exceptionally well informed
and never tempted to abuse his power. However, thought experi-
ments are processes that we can approach slowly and reflectively,
thereby guarding against possible biases. If such biases occur, this
does not rule out the use of thought experiments, but rather requires
us to redesign them, especially as similar, if not greater, biases are
likely to plague actual, real-world scenarios.

4 Why Wacky Thought Experiments Can Be Well-Posed

We reject the possibility that a bar on wackiness is one condition of a well-
posed thought experiment. In this section, we explain why.
The ‘wackiness’ of thought experiments can be disambiguated into the
two main categories noted in Sections 1 and 2: imaginative opacity and
necessary hypotheticality, including necessary hypotheticality for us here and
now. We hold that neither of these dimensions of wackiness bars thought
38 Chapter 3

experiments from being well-posed. Before we consider imaginative opa-

city or necessary hypotheticality, we need to address an objection that
applies not only to these dimensions of wackiness, but also to something
that is not at all wacky: contingent hypotheticality.

4.1 Contingently Hypothetical Thought Experiments

Some critics seem to hold that we can conduct respectable normative
theory without thought experiments. Such critics maintain that it does
not matter that a given theory or principle would lead to counterintuitive
recommendations when checked against a hypothetical scenario because
all we need to know is whether our principles will serve us in our actual
world (e.g. Hare 1981: 135 and chapter 8 passim; for discussion, see
Carson 1993). The suggestion is that all that matters in normative theory
is whether our principles perform well in the actual world, and not how
well they cope with non-actual situations.
Against this view, we offer two arguments. First, the performance of our
principles in actual scenarios is not all that matters in normative theory.
Frequently, normative principles are put forward as explanations of why
a given course of action is right or wrong. Call such principles explanatory
principles. Explanatory principles must be more abstract than the principles
that bear on a given actual situation if they are to explain our judgements in
that situation. For instance, it is wrong to kill ﬁve patients in order to harvest
their organs to save one patient because it is wrong to kill in order to save.
Since such explanatory principles are more general than the situations that
call for them in the actual world, they must be testable against all situations
to which they could, in principle, apply. This means that they also stand or
fall by their performance in hypothetical situations. This is one reason that
political theorists often need to consider counterexamples, and rely on
hypothetical scenarios to expand their range, in order to challenge abstract
principles.
Second, even if we are not seeking explanatory principles, but merely
principles that will serve us well with the problem at hand, thought
experiments offer us a way of trying to resolve disagreement on which
principles these are. All else being equal, it should count in favour of
a principle that it holds across a wider range of scenarios (just as it counts
in favour of a theory that it applies to a wider range of problems). So,
thought experiments are important in normative theory because they can
help us break ties.
These responses might be taken to suggest that normative theories
(call them applied theories) would not need thought experiments if
they did not seek explanatory principles or were not challenged by
Thought Experiments 39

other theories. But even this suggestion should be resisted. That is,
even applied theories that carry support need a method to discern
whether the considerations brought to bear on a given problem (to
reach an all-things-considered judgement) are relevant from the moral
point of view. For example, we need to ask whether it is relevant to
consider such things as ‘this is my house,’ or ‘this is extremely
demanding,’ or ‘she is a woman,’ or ‘it’s repellent,’ or ‘I would not
do it myself.’ Thought experiments offer an efﬁcient way of testing the
salience of such considerations (which is not to say that such testing
would amount to a proof. On the fact that thought experiments alone
do not deliver transparent data, see Kagan 1988).

4.2 Imaginative Opacity

Turning to imaginative opacity, this dimension of wackiness raises the
following worry (Cooper 2005: 328–47; Parﬁt 1986: 389; Wilkes 1988:
15ff). Imaginatively opaque thought experiments fail to have an adequate
imaginative grip, and hence they pose ‘what if?’ questions that the experi-
menter cannot answer. The reason that the experimenter cannot answer
those questions may be that she has no knowledge of the laws that govern
the behaviour of the entity she is imagining. Or she may have knowledge
of the laws relevant for predicting that behaviour in the actual world (e.g.
the process of human birth), but those laws do not apply in the hypothe-
tical scenario (e.g. giving birth to oneself). The fact that the experimenter
cannot answer these questions is said to negate whatever argumentative
value the thought experiment might have.
At least two replies can be made to this objection. First, ruling out the
use of imaginatively opaque thought experiments would be unduly pro-
hibitive. It would rule out the use of thought experiments that expose
certain paradoxes. For instance, it would rule out the use of a thought
experiment used to show that causal paradoxes would emerge if one could
go back in time and kill one’s father. But the opacity of the experiment
does not stop us from pointing out the potential paradoxes.
Second, it is not clear that being able to imagine all aspects of a given case
is essential to run the thought experiment. For example, we (the authors of
this chapter) cannot fully conceive of a being that is both a dog and able to
talk, but we can still ask whether such a dog would count as a person.
Likewise, we cannot imagine a utility monster that derives almost bound-
less pleasure from the suffering of others, but we can ask whether such
a being would be right to make others suffer. Recall that thought experi-
ments are not ‘run’ (simply) to establish how we (the experimenters) would
feel, but to establish what we may plausibly think, and hence we may not
40 Chapter 3

require a full character brief in the way that actors would if they were
required to play a given part and to react ‘in character’.

4.3 Necessary Hypotheticals

The worry that we do not understand the laws that govern some imagina-
tively opaque cases resurfaces in a form that applies also to necessarily
hypothetical thought experiments. The worry takes the form of
a dilemma:
Thought experiments are useless because we either cannot set them up
properly or cannot derive any credible conclusions from them. That is,
either we are assuming a world similar to ours, in which case we cannot
set up a wacky thought experiment at all (e.g. in a world similar to ours,
dogs are not as intelligent as persons; there is no teleporting; etc.), or we are
assuming a world that is radically different from ours, in which case we
cannot know what to say about this world. (See Raz 1986: 419–20; Mulhall
2002: 16–18.)
Why can we apparently not know what to say about this world?
The answer relates to ‘semantic holism’ (see Sorensen 1992: 282–4).
The idea is that our concepts developed to track our world, rather than
the wacky worlds that we set up in our experiments, and wacky worlds
cannot plausibly capture our real-life concepts. For instance, suppose
that half of the standard tests that we use to determine whether a piece
of copper wire is electriﬁed give us a positive answer and half of the tests
give us a negative answer. What should we say about this piece of copper
wire? The answer, according to Nowell-Smith, should be: ‘I simply do not
know what I should say’ (1954: 240; quoted in Sorensen 1992: 283).
Now consider a wacky, normative-theory thought experiment (inspired
by Roger Crisp’s Beverly Hills Case; Crisp 2003: 755).
Rich and Superrich: Imagine a world in which there are only rich and superrich.
Is the inequality that holds between the rich and the superrich unfair or otherwise
problematic?

The answer, according to a critic of wacky thought experiments, is

‘I simply do not know’ since our concepts developed to deal with entirely
different cases and they are of no use in radically reimagined worlds.
To give an analogy, paint colours developed to paint the British landscape
are of little use in painting the African landscape, given the very different
light of the two environments.
However, this objection rests on a mistake. It assumes that thought
experiments ask us what we would say if our concepts were developed to
accommodate the wacky cases as standard. But this is not what thought
Thought Experiments 41

experiments ask us to do. They ask us, instead, to judge how our current,
familiar concepts behave when exposed to new situations. To see this,
first return to the paint analogy. The thought here is that we are not asked
to use the British paints to paint the African landscape; we are asked
instead to use the African light to rule, say, on whether two identical-
looking British colours really are identical. When we cannot easily tell if
the colours are the same against the British light, we may benefit from
examining them under the African light. We examine the value of equality
by asking if we still value equality in such a context. If not, then we have
reason to suspect that what matters to us in our ordinary context is not
simply equality, but absolute levels of deprivation.
Second, consider another illustration of the application of our ordinary
concepts to new situations. In Rich and Superrich, we do not ask what we
would think if we were living in such a privileged society. Instead, we ask
whether we consider the inequality present in that privileged society too
unfair by our own, current standards. Pointing out that in a world with
only the rich and the superrich, no one would care about equality (and
that they may not even have a sense that they are unequal) is irrelevant to
the question of whether we now see the inequality as problematic. (There
is a wrinkle here. We may have a conception of unfair inequality according
to which inequality is only unfair if the people subject to it consider it
unfair; if this is so then, indeed, we may be unable to tell whether unfair
inequality characterizes the hypothetical scenario, but that is not because
the thought experiment is hypothetical and wacky, but because we do not
have the relevant empirical data about the people we are investigating.)
To conclude this section, we want to emphasize that wacky thought
experiments are not, in fact, used solely to advance academic debates,
which some might consider esoteric in any case. Testing how our
familiar concepts behave when exposed to new situations is
a common, undisputed, and powerful strategy in many fields, includ-
ing one closely aligned with normative theory, namely, law. It is
routine practice in law schools to hold moot courts revolving around
wild and wonderful cases so as to train law students in the application
of key legal concepts. And the application of such legal concepts as
theft and property is viewed as no less legitimate when the parties are
Martians stealing magic from Venusians. Moreover, in broad terms,
real court cases are exercises in thought experimenting since both
ordinary court procedures and norms of due process necessarily
yield an abstracted and idealized presentation of the facts of the case.
Ultimately, wacky thought experiments are not undermined by our
inability either to imagine all of their elements or to anticipate how the
concepts we are exploring would evolve in hypothesized worlds.
42 Chapter 3

5 Conclusion: How to Design Good Thought Experiments in

Political Theory
This discussion has identiﬁed several conditions for good thought
experiments. We conclude here by presenting these conditions as a set
of ‘how-to’ instructions for the design of good thought experiments in
normative theory:
1. We may stumble across thought experiments in ﬁlm and novels; and
life itself can generate situations that can be thought through as
thought experiments. But, usually the impetus for designing them is
that we want to test some premise in our argument/some hypothesis
about a plausible principle or value. If so, the most fundamental
question you should pose is this: what exactly would you like
your thought experiment to test (i.e. to undermine or to
support)?
2. Ensure that the thought experiment is relevant to what is being
tested (see Section 3.2). Try to design the simplest experiment in
the sense of having the most parsimonious story (to avoid introducing
distorting elements), but do not worry if the story is also wacky in the
sense of being fantastical. (Our defence of wacky thought experiments
is meant to set your imagination free.)
3. If your thought experiment involves a denial of a standard
feature of the world (e.g. you deny that the police are uncertain
whether they caught the right guy), hypothesize, even roughly,
how this can be – what alternative feature of the world is present. (See
Section 3.2.1.)
4. Be sensitive to possible narrative-framing biases, that is, to the
structure of the scenario and to the narrative assumptions it can imply.
(See Section 3.2.2.)
5. Consider whether any imaginatively opaque elements of the
thought experiment need to be made imaginatively clear in
order for the thought experiment to function as intended. (See
Section 4.2.)
6. When constructing a necessary hypothetical, consider whether the
selected ‘foreign context’ best serves to illuminate the features
of our ordinary concepts and/or principles that are under con-
sideration. (See Section 4.3.)
7. Ensure that none of the features of the experiment already
assumes what is to be tested. (See Section 3.1.1.)
8. Ensure that the thought experiment, together with the hypoth-
esis it is meant to support or deny, translates into a valid argu-
ment. (See Section 3.1.2.)
Thought Experiments 43

Appendix: Further Examples of Thought Experiments

1. Judith Jarvis Thomson’s Famous Violinist: A healthy person awakes
in a hospital to find that, unbeknownst to her, she has been connected
by her kidneys to a famous violinist. The famous violinist will die unless
she remains connected to him for the next nine months, just as a foetus
will die unless it remains ‘connected’ for nine months to the pregnant
woman carrying it. Is the healthy person morally obligated to remain in
the hospital with the violinist for nine months? (Thomson 1971: 48–9).
The thought experiment is used to challenge the impermissibility of
abortion.
2. Judith Jarvis Thomson’s Loop Trolley: The two tracks in the Trolley
Problem split, but then circle back to form a loop. Diverting the trolley
onto the track with the one person will cause the trolley to hit the one
and thereby prevent the trolley from continuing around the loop to hit
the five. Is it permissible to turn the trolley away from the track with the
five and onto the track with the one? (Thomson 1985: 1402–3).
This thought experiment is used to show that, like in the Transplant Case,
turning the trolley uses the one person merely as a means.
3. Judith Jarvis Thomson’s Fat Man on the Bridge: The runaway
trolley of the Trolley Problem cannot be stopped by diverting it, but
can be stopped by pushing a fat man off a bridge onto the track in front
of the trolley, resulting in his death. Is it permissible to push the fat
man? (Thomson 1985: 1409).
This thought experiment is used to show that pushing the fat man uses
him merely as a means and infringes his right not to be so treated.
4. Bernard Williams’s Jim and the Indians: A jungle explorer, Jim,
comes across a colonialist, Pedro, who has twenty aboriginal people
lined up before a firing squad. Pedro offers Jim the ‘privilege’ of killing
one and letting the rest go free. Jim can either refuse the offer, in which
case all twenty people will be killed, or he can accept the offer and kill
only one of them himself. What should Jim do and how should he feel
about his choice? (Williams 1973: 98–100).
This thought experiment is used to question the thought that a plausible
moral theory must be consequentialist. Focusing only on the lives lost, it
seems that Jim must kill the one; but Williams argues that morality is not
only about good and bad states of affairs.
5. Robert Nozick’s Wilt Chamberlain: Suppose that basketball
player Wilt Chamberlain attracts huge crowds to his games. He
signs a contract with a team whereby he receives twenty-five cents
for every home game ticket sold. During one season, a million
people come to his games and happily drop twenty-five cents into
44 Chapter 3

a special box for him, thinking it well worth it. He earns $250,000
this way, a much larger sum than the average income. Is he entitled
to keep all of it?
This thought experiment is used to show that people’s exercises of
personal freedom, such as choosing to pay twenty-ﬁve cents to go to
Chamberlain’s games, will disrupt a patterned principle of distributive
justice, such as the principle that everyone should have equal resources.

Acknowledgements
We thank Adrian Blau, Cécile Fabre, Fay Niker, Jonathan Quong, Victor
Tadros, and Lea Ypi for their very helpful feedback on this chapter.
We also thank Fay Niker for her research assistance. We thank the
participants of the Oxford Political Theory Seminar (June 2010) and
the Manchester Centre for Political Theory Seminar (October 2010) for
their comments.

References
Brown, James Robert, 1986. ‘Thought experiments in the scientiﬁc revolution’,
International Studies in the Philosophy of Science 1: 1–15.
Brown, James Robert, and Fehige, Yiftach, 2014. ‘Thought experiments’, in
Edward Zalta (ed.), The Stanford Encyclopedia of Philosophy (Spring 2016
Edition). https://plato.stanford.edu/archives/spr2016/entries/thought-experiment/.
Carson, Tom, 1993. ‘Hare on utilitarianism and intuitive morality’, Erkenntnis 39:
305–31.
Cohen, G. A., 2009. Why Not Socialism? Princeton, NJ: Princeton University
Press.
Coleman, Stephen, 2000. ‘Thought experiments and personal identity’,
Philosophical Studies 98: 53–69.
Cooper, Rachel, 2005. ‘Thought experiments’, Metaphilosophy 36: 328–47.
Crisp, Roger, 2003. ‘Equality, Priority, and Compassion’, Ethics 113: 745–63.
Foot, Philippa, 2002. ‘The problem of abortion and the doctrine of double effect’,
in Virtues and Vices and Other Essays in Moral Philosophy. Oxford: Clarendon
Press, 19–32.
Gendler, Tamar Szabó, 1998. ‘Galileo and the indispensability of scientiﬁc
thought experiment’, British Journal for the Philosophy of Science 49: 397–424.
Haggqvist, Sören, 1996. Thought Experiments in Philosophy. Stockholm: Almqvist &
Wiksell International.
Hare, R. M., 1981. Moral Thinking. Oxford University Press.
Kagan, Shelly, 1988. ‘The Additive Fallacy’, Ethics 99: 5–31.
Martin, C. B., 1958. ‘Identity and exact similarity’, Analysis 18: 83–7.
Mulhall, Stephen, 2002. ‘Fearful thoughts’, The London Review of Books 24 (16):
16–18.
Thought Experiments 45

Norton, John, 1996. ‘Are thought experiments just what you thought?’, Canadian
Journal of Philosophy 26: 333–66.
Nowell-Smith, P. H., 1954. Ethics. London: Penguin Books.
Nozick, Robert, 1974. Anarchy, State, and Utopia. Oxford: Blackwell.
O’Neill, Onora, 1986. ‘The power of example’, Philosophy 61 (235): 5–29.
Parfit, Derek, 1986. Reasons and Persons. Oxford University Press.
Putnam, Hilary, 1975. ‘The meaning of “meaning”’, in Keith Gunderson (ed.),
Language, Mind, and Knowledge. Minneapolis: University of Minnesota Press,
131–93.
Rawls, John, 1971. A Theory of Justice. Cambridge, MA: Harvard University Press.
Raz, Joseph, 1986. The Morality of Freedom. Oxford University Press
Rivera-Lopez, Eduardo, 2005. ‘Use and misuse of examples in normative ethics’,
The Journal of Value Inquiry 39: 115–25.
Singer, Peter, 1972. ‘Famine, affluence, and morality’, Philosophy and Public Affairs
1: 229–43.
Sorensen, Roy, 1992. Thought Experiments. Oxford University Press.
Thomson, Judith Jarvis, 1971. ‘A defense of abortion’, Philosophy and Public Affairs
1: 47–66.
Thomson, Judith Jarvis, 1985. ‘The trolley problem’, Yale Law Journal 94:
1395–1415.
Unger, Peter, 1982. ‘Toward a Psychology of Common Sense’, American
Philosophical Quarterly 19: 117–129.
Walzer, Michael, 1973, ‘The problem of dirty hands’, Philosophy and Public Affairs
2: 160–80.
Williams, Bernard, 1960. ‘Bodily continuity and personal identity’, Analysis 20:
117–20.
Williams, Bernard, 1973. ‘A critique of utilitarianism’, in J. J. C. Smart and
Bernard Williams (eds.), Utilitarianism: For and Against. Cambridge University
Press, 75–150.
Wilkes, Kathleen, 1988. Real People: Personal Identity without Thought Experiments.
Oxford: Clarendon Press.
Winch, Peter, 1965. ‘The universalizability of moral judgments’, The Monist 49:
196–214.
Wittgenstein, Ludwig, 1965. ‘A lecture on ethics’, The Philosophical Review
74: 3–12.
4 Reflective Equilibrium

Carl Knight

1 Introduction
The method of reflective equilibrium focuses on the relationship between
principles and judgments. Principles are relatively general rules for compre-
hending the area of enquiry. Judgments are our intuitions or commit-
ments, ‘at all levels of generality’ (Rawls 1975: 8), regarding the subject
matter. The basic idea of reflective equilibrium is to bring principles and
judgments into accord. This can be achieved by revising the principles
and/or the judgments. For instance, if I am considering the principle that
it is always wrong to lie, but have the judgment that it would not be wrong
to lie in order to save a life, I can reach equilibrium by revising either the
principle or the judgment.
Reflective equilibrium is the most widely used methodology in con-
temporary moral and political philosophy (Sinnott-Armstrong et al.
2010: 246; Varner 2012: 11). It has even been suggested that it is ‘the
only defensible method’ (Scanlon 2003: 149). Its popularity is undoubt-
edly strongly influenced by John Rawls’s use of it in his seminal A Theory
of Justice, published in 1971.1 However, the method precedes this, and
extends to other fields. For instance, Nelson Goodman wrote regarding
induction that ‘[t]he process of justification is the delicate one of making
mutual adjustments between rules and accepted inferences; and in the
agreement reached lies the only justification needed for either’ (Goodman
1965: 64). Some fields, by contrast, do not seem as amenable to the
method of reflective equilibrium. Within linguistics, for instance, native
speakers’ judgments of grammaticality cannot generally be replaced as
moral judgments can (Daniels 1996: ch. 4). Although most writers
treat reflective equilibrium as unproblematic within empirical sciences
(Daniels 1996: 31–3; Cummins 1998; Welch 2014: 4; see also
McDermott 2008), adjustment of empirical judgments also seems to be

1
Jo Wolff (2013: 808) notes that, of the papers collected in the ﬁrst two series of Politics,
Philosophy, and Society, Rawls’s was unique in aiming to defend a substantive position, and
in deploying a distinctive methodology to positive effect.

46
Reﬂective Equilibrium 47

subject to stronger constraints than those that apply to moral judgments

(see Singer 2005: 345).
Indeed, here there seems to be a significant difference between norma-
tive political theory and empirical political science. A normative political
theorist who, to her surprise, finds that a confidently held moral judgment
conflicts with an otherwise compelling principle (or set of principles) is
free to reject that judgment precisely because it conflicts with the favoured
principle. But it would be quite improper for a confidently held empirical
judgment to be abandoned simply because it turned out not to fit with the
investigator’s pet hypothesis. Full reflective equilibrium, with judgments
adjusted at will just as principles are, is primarily the reserve of normative
political theory. It is with normative political theory that this chapter will
be concerned.
I first look at normative political judgments (Section 2) before consid-
ering the role of principles, arguments, devices of representation and
background theory in wide reflective equilibrium (Section 3). I then
consider two of the main challenges to the method (Section 4), and
show how to use it to deliberate about substantive political principles
(Section 5). I conclude with an extended example of the method in action
(Section 6).

2 Judgments
The starting point for reflective equilibrium is our judgments. We cannot,
however, use just any judgments. For instance, judgments made ‘in the
heat of the moment’ would not be a reliable basis for equilibrium.
Considered judgments are what we need. These are ‘those judgments in
which our moral capacities are most likely to be displayed without distor-
tion’ (Rawls 1999: 42).
Most writers, including Rawls (1999: 42), suppose that, in order to
count as considered, judgments should be held with confidence. Indeed,
Rawls often uses ‘convictions’ as a synonym for ‘judgments’ (Rawls 1975:
8; 2005: 24, 26, 28, 151, 156). This ‘confidence constraint’ seems to me
quite gratuitous (Knight 2006: 207–8). If I have the firm conviction that
the state should protect its citizens from terrorism, and also believe, less
firmly, that individuals have a right to privacy, a right to not be subject to
pre-trial detention beyond a certain duration and a right to a public trial,
the confidence constraint would seem to require that my numerous but
less firmly held concerns about individual rights be set aside. But this is to
give free reign to the one firm conviction, with the upshot that the
principle(s) arrived at in reflective equilibrium will allow almost any
breach of civil rights in the name of public protection. This is in
48 Chapter 4

contradiction of the majority of the judgments I hold and (most likely) my

overall view.
Undoubtedly, a firmly held judgment should generally carry more
weight in our deliberations than a less firmly held one. But reflective
equilibrium automatically does that, as we are presumably more likely to
give up our less firmly held judgments in the face of opposing judgments
or principles. If we do not, that is because it turns out that the less firmly
held judgments had something going for them. Maybe they individually
or collectively capture something that, on reflection, we consider impor-
tant. Thus, I think we should reject the confidence constraint.
A different constraint has sometimes been assumed, including in my
earlier work (Knight 2006: 207). This specifies that our considered judg-
ments do not display errors of reasoning, such as logical inconsistencies or
empirical errors. Some writers go further, suggesting that we should
disregard judgments that we do not have evidence for, as they are not
epistemic assets (Gaus 1996: 86; Kelly and McGrath 2010: 347–4).
Should we, then, endorse an ‘epistemic constraint’, requiring that only
justified or warranted judgments, or (more minimally) only those that
lack errors, are admitted to the reflective equilibrium procedure?
Though this may seem like simple common sense, I doubt it. Consider
first the stronger version of the constraint, which requires justification or
warrant for a judgment. Evidence can be rather thin on the ground when
we are dealing even with firmly held judgments. If I consider some
political judgment that I hold very firmly, such as the judgment that no
fellow citizen should avoidably starve, it is hard to point to anything that
can really count as evidence for that belief in the relevant sense. I might
point out that my compatriot will suffer pain, reduced capability and
eventually death, but these empirical facts alone cannot really be evidence
for the normative judgment I am making. It seems that, in a case like this,
the judgment itself is foundational. My judgment seems pre-theoretically
plausible to me, and that is sufficient to grant it ‘independent credibility’
(Hooker 2012: 23). This does not mean that it has any weight in my final
principles, but it is enough for it to be granted admission to the reflective
equilibrium process. It is there that the credentials of our judgments are
really tested, by seeing how well they fit with our other judgments and the
most plausible principles, in light of the most compelling arguments we
can muster.
What then of the less demanding version of the epistemic constraint,
which requires merely the absence of outright error? Surely we can reject
some judgments as clearly erroneous. But even this constraint might be
thought to be excessively demanding in that it goes beyond providing
‘conditions favorable for deliberation and judgment in general’ (Rawls
Reflective Equilibrium 49

1999: 42) and actually limits the admissible content of judgments.

Furthermore, exactly what qualifies as an error of reasoning and what
qualifies as an empirical error is controversial. We could consider these
issues in piecemeal fashion prior to entering reflective equilibrium. But
this is counterproductive as we have no way of knowing whether these
isolated speculations will be consistent with the most plausible overall
position. We should instead consider these issues holistically, as pieces in
the jigsaw that is the coherent view of the conceptual terrain that we aim
to arrive at in reflective equilibrium. (Specifically, these issues are settled
through consideration of relevant background theories – see Section 3.)
Reflective equilibrium eschews the essentialist notion ‘that we can deter-
mine the nature of certain facets of these inquiries in advance of the
inquiries themselves, and that nothing that comes about in inquiry will
change those facets’ (Walden 2013: 255). The epistemic constraint, even
in its minimal form, seems to put the cart before the horse, and should be
discarded. Considered judgments are just those made in ‘conditions
favorable for deliberation and judgment in general’.

3 Wide Reﬂective Equilibrium

Suppose that you have arrived at your set of considered judgments.
You might first use these to reach narrow reflective equilibrium, in
which ‘one is to be presented with only those descriptions which more
or less match one’s existing judgments except for minor discrepancies’
(Rawls 1999: 43; see also Rawls 2005: 8 n. 8). Narrow reflective
equilibrium is in essence an effort to systematize an agent’s pre-
theoretical views. As such, it has limited epistemic value. Were some-
one to ask you what justification you have for your principles, you do
not have much of a reply. To be sure, the narrow reflective equili-
brium principles might be an improvement from your perspective on
the bare intuitions you started out with. But you can hardly say that
your principles are well justified where they are just a direct expression
of your pre-theoretical intuitions.
The more interesting version of the method is wide reflective equili-
brium. Rawls describes this in very demanding terms: ‘one is to be
presented with all possible descriptions to which one might plausibly
conform one’s judgments together with all relevant philosophical argu-
ments’ (Rawls 1999: 43). So for wide reflective equilibrium to be reached,
you must consider all principles (and combinations of principles) that you
might accept. As one way in which you may conform your judgments to
principles is to change your judgments, this means that you must consider
every principle in every combination with every other principle!
50 Chapter 4

Unsurprisingly, Rawls does not attempt to fully satisfy this unachieva-

ble standard, resolving in A Theory of Justice to compare only his own
‘principles and arguments with a few other familiar views’ (Rawls 1999:
43). For all practical purposes, it will undoubtedly be necessary to narrow
our equilibrium in this way. Nevertheless, I think there is great value in
keeping in mind that wide reflective equilibrium is an ideal. It sets the bar
high. Though the theorist will inevitably only consider a few principles,
this is not because that is all the method of reflective equilibrium requires
for a full justification to be provided. It should always be kept in mind that
consideration of more principles would provide a fuller justification.
Furthermore, if we have to cut corners, we should do so in the way least
harmful to the strength of the final justification. This means, for example,
ensuring that we at least consider the most compelling rival principles,
rather than satisfying ourselves with seeing off straw men.
Reflective equilibrium can be interpreted as providing an ecumenical
answer to a long-standing problem in epistemology. The ordinary way of
justifying beliefs is inferential and linear: belief A justifies belief B, which
justifies belief C, and so on. The problem here is rather obvious. As the
chain of inference cannot go on infinitely, it seems that none of our beliefs
will be justified. There are two ways out of this infinite justificatory
regress. Foundationalism denies that all justification is inferential – for
example, A might be justified by something other than another belief.
Coherentism denies that all justification is linear – for example, C might be
justified by A (Brink 1989: 109). A large majority of writers see reflective
equilibrium as a coherentist method (Brink 1989: 134; Daniels 1996:
60–1; Tersman 2008: 398–400; Maffetone 2010: 142–5), while a few see
it as foundationalist (DePaul 1986; Ebertz 1993). In my view it clearly
contains elements of both approaches. Foundationalism can be seen to be
present as I would, according to the method, be justified in favouring one
possible coherent set of principles and considered judgments to another
purely because the former coincides with my actual considered judg-
ments. Coherentism cannot explain this, as each set is identical as regards
coherentist non-linear justificatory chains. But coherentism is evidently
also present, as the method says that the fact that some judgment (or
principle) coheres with the rest of our beliefs counts in its favour.
As I have mentioned, Rawls requires not only that principles be con-
sidered, but relevant arguments as well. We could reach equilibrium
without arguments, but coherence among beliefs that have not been
subjected to serious scrutiny would be of limited justificatory value.
This introduces several new complexities. First, there are arguments
that directly support or undermine judgments and principles. For
instance, when contemplating utilitarianism, the objection that
Reflective Equilibrium 51

Table 4.1: Elements of wide reﬂective equilibrium

Element Scope Role Examples

Judgments Speciﬁc or general Primary subject of Racial discrimination is

equilibrium wrong; all individuals
have equal moral
worth.
Principles General Primary subject of The difference
equilibrium principle; equal
moral worth principle
Direct arguments Speciﬁc Argumentation Rawls’s intuitive
argument; the
levelling down
objection
Devices of General Argumentation The original position;
representation the ideal observer
Background theories General Argumentation Theories of the person;
social theory

utilitarianism seems in some circumstances to permit slavery or knowing

punishment of the innocent should be considered (Varner 2012:
11). Second, there are structures for framing our deliberations that go
beyond single arguments, which Rawls terms ‘devices of representation’
(Rawls 2005: 23–8). These typically provide special circumstances for
principle selection, with the parameters of those circumstances set by the
theorist’s judgments regarding what is reasonable or rational. Rawls’s
original position is the best known example within analytical political
philosophy. There are many more examples in contemporary work
(Ackerman 1980; Gauthier 1986; Dworkin 2000: ch. 2) and, arguably,
older social contract theory (Hobbes, Locke) and ideal observer theory
(Hume, Smith). Finally, there are ‘background theories’ (Daniels 1996:
22–3), which are drawn on by both the direct arguments and the devices
of representation, and themselves tested for intuitive appeal. For
instance, if a theory of the separateness of persons were found compelling,
it might be used to undermine certain principles, as Rawls (1999: 23–4)
seems to argue is true of utilitarianism. The various elements of wide
reﬂective equilibrium are summarized in Table 4.1.
In practice, it may not always be easy to distinguish the different
elements, and it is not absolutely essential to do so. For instance, the
table gives an example of a principle (equal moral worth principle) that is
more or less a restatement of a judgment (all individuals have equal moral
worth). As judgments and principles, qua judgments and principles, do
52 Chapter 4

not receive privileged epistemic status – ‘[o]ur “intuitions” are simply

opinions: our philosophical theories are the same’ (Lewis 1983: x; see also
Freeman 2007: 33; Mandle 2009: 171–2) – there is no problem with the
boundaries between them being fuzzy or overlapping.2 Judgments and
principles are only distinguished here as this is a familiar and often helpful
way of arranging our thoughts. Likewise, and as indicated in the table, the
direct arguments, devices of representation and background theories are
really just subsets of one big category of ‘argumentation’. They do not
need to be systematically separated as none has priority over any other.
Some of the argumentation elements may even be absent in the creation
of particular equilibria; for instance, the extended example in Section 6
does not refer to device of representation.

4 Challenges
As the most widespread approach to theory selection in moral and poli-
tical philosophy, the method of reflective equilibrium has faced its share
of critical attention. In this section I consider a couple of the more
significant challenges.
A common complaint with the method is that it relies entirely on the
quality of the judgments that form a central part of the equilibrium
(Brandt 1979: 20; Williamson 2007: 244–6). Advocates of the method
typically build their examples around highly plausible judgments, such as
Rawls’ convictions about the wrongness of religious intolerance and racial
subordination. But if someone starts with implausible or even repugnant
judgments, there is, critics claim, nothing to stop the method from gen-
erating implausible conclusions. The point is put clearly by Thomas Kelly
and Sarah McGrath (2010: 346–7):
It is a good objection to a method if it turns out that impeccably following that
method could lead one to views that are unreasonable. It follows from this that if
beginning from all and only one’s considered judgments, and from there achiev-
ing wide reflective equilibrium without making any ‘downstream’ mistakes, is
sufficient for impeccably executing the method of reflective equilibrium, then the
method is not correct. The problem is that something might very well qualify as
a considered judgment, when that notion is understood in anything like the way it
is understood in the broadly Rawlsian tradition, and yet be utterly lacking in
rational credibility.
This is illustrated with the observation that there is nothing to stop ‘[o]ne
is morally required to occasionally kill randomly’ from counting as

2
Welch even defends a radical version of reﬂective equilibrium in which ‘there are no
considered judgments to consider’ (Welch 2014: 14).
Reﬂective Equilibrium 53

a considered judgment. Kelly and McGrath therefore conclude that

reflective equilibrium is an inadequate method.
This critique seems to be misdirected in several respects. First, Kelly
and McGrath focus on considered judgments to such a degree that
reflective equilibrium proper falls out of their picture entirely. They
seem to take it as given that the final set of principles will simply be direct
expressions of the initial considered judgments. While that may be more
or less true of narrow reflective equilibrium, it is unlikely to be true of wide
reflective equilibrium. Sustained consideration of competing principles,
supporting arguments, devices of representation and background theories
is extremely likely to expunge judgments that are ‘utterly lacking in
rational credibility’,3 in which case the alleged problem does not arise.
Second, it is not clear that we have actually been shown a case in which
‘impeccably following’ the method of reflective equilibrium ‘lead[s] one
to views that are unreasonable’. In Kelly and McGrath’s example, the
random killing judgment is held initially. So it is not the case that the
method ‘leads’ anyone to this judgment. Rather, they had the judgment to
begin with. If there is a complaint to be had here, it is with the life history
that has resulted in such an absurd judgment being formed.
Finally, I doubt that it actually is ‘a good objection to a method if
it turns out that impeccably following that method could lead one to
views that are unreasonable’. Kelly and McGrath (2010: 327–8)
support this claim with the following example:
Suppose that, prior to embarking upon the systematic study of fruit flies, one held
various baseless opinions about their nature. If one then devoted oneself to the
study of fruit flies, and impeccably followed the best scientific procedures we have
for arriving at accurate views about their nature, we would expect those earlier
baseless opinions to be filtered out or corrected at some stage in the inquiry. In the
unlikely event that some of those opinions were among the views that one held
after . . . impeccably following our best scientific methods, then, we submit, those
beliefs would no longer be unreasonable ones to hold.
The conclusion may seem plausible here on account of misleading fea-
tures of the case. In particular, the ‘baseless opinions’ are so sparsely
described that we have no way of grasping whether they might be held
reasonably or not. To really test the central underlying claim here that
application of the scientific method, unlike reflective equilibrium,
removes unreasonable beliefs, we should adjust the scientific baseless
opinions, so that they are as vivid as their moral counterpart – the judg-
ment that ‘[o]ne is morally required to occasionally kill randomly’.

3
Kelly and McGrath seem to concede a similar point regarding empirical sciences – see the
lengthy quote given two paragraphs later.
54 Chapter 4

So suppose that the baseless beliefs about fruit flies are the following: fruit
flies originate from specific acts of divine creation; these acts occurred
within the past 10,000 years and are described in scripture; it is a matter of
religious duty to disregard all countervailing evidence regarding the ori-
gins of fruit flies. I think it highly plausible that these views are unreason-
able, and that ‘devot[ing] oneself to the study of fruit flies, and
impeccably follow[ing] the best scientific procedures we have for arriving
at accurate views about their nature’ does not stop these views from being
unreasonable. The lesson to draw from this is that neither the method of
reflective equilibrium nor the scientific method is guaranteed to rid people
of unreasonable beliefs. But that doesn’t change the fact that both are
more likely than alternatives to provide individuals with reasonable
beliefs, by exposing them to the most compelling evidence available in
their respective fields.
This leads us to the second challenge. Several writers have claimed
not that reflective equilibrium struggles with implausible idiosyncratic
judgments, like the random killing judgment, but with the fact that
our judgments are systematically undermined (Brandt 1979: 21–2;
Hare 1981: 12). Peter Singer emphasizes that our moral judgments
have largely arisen through an evolutionary process. For example, the
commonsense idea that we have stronger duties towards relatives can
be explained on the basis that the corresponding genes ‘are more likely
to survive and spread among social mammals than genes that do not
lead to preferences for one’s relatives that are typically proportional to
the proximity of the relationship’ (Singer 2005: 334; see also Singer
1974). It is no surprise, then, that brain scans suggest that our moral
judgments often do not seem to be informed by reason, but are rather
an immediate emotional response (Singer 2005: 339–42). Individuals
will stick to their judgment even where they end up rejecting the
reasons they initially give for it (Singer 2005: 337–8). This modern
scientific understanding of ‘how we make moral judgments casts ser-
ious doubt on the method of reflective equilibrium’, according to
Singer (2005: 348):
There is little point in constructing a moral theory designed to match considered
moral judgments that themselves stem from our evolved responses to the situa-
tions in which we and our ancestors lived during the period of our evolution as
social mammals, primates, and finally, human beings. We should, with our
current powers of reasoning and our rapidly changing circumstances, be able to
do better than that.
Suppose, for the sake of argument, that the evolutionary picture that
Singer paints is correct. I would not see this as a threat to the method of
Reflective Equilibrium 55

reﬂective equilibrium. Singer is, in effect, presenting a background theory

that should be considered when an individual is undergoing reflective
equilibrium.4 If the background theory is compelling, as I suspect it might
be, that may cause individuals to treat their moral judgments differently,
taking care to consider whether a judgment might amount to an evolved
emotional response that should be set aside.
Singer (2005: 347) anticipates a response along these lines, and replies
as follows:
Admittedly, it is possible to interpret the model of reflective equilibrium so that it
takes into account any grounds for objecting to our intuitions, including those that
I have put forward. Norman Daniels has argued persuasively for this ‘wide’
interpretation of reflective equilibrium. If the interpretation is truly wide enough
to countenance the rejection of all our ordinary moral beliefs, then I have no
objection to it. The price for avoiding the inbuilt conservatism of the narrow
interpretation, however, is that reflective equilibrium ceases to be a distinctive
method of doing normative ethics. Where previously there was a contrast between
the method of reflective equilibrium and ‘foundationalist’ attempts to build an
ethical system outward from some indubitable starting point, now foundational-
ism simply becomes the limiting case of a wide reflective equilibrium.

Here Singer claims that reflective equilibrium would have to rely not just
on the moderate, revisable foundationalism referred to earlier, but rather
on a stronger ‘special foundationalism’ (Harman 2003: 415) that identi-
fies certain ethical truths as unchallengeable. Were that true, it would
certainly be the case that reflective equilibrium had been stripped of its
distinctive features (in particular, mutual adjustment of judgments and
principles). But it is not true. Singer says that the interpretation should be
‘truly wide enough to countenance the rejection of all our ordinary moral
beliefs’. Reflective equilibrium is this wide (Sandberg and Juth 2011:
222). However, Singer’s conclusion implicitly assumes that countenan-
cing the rejection of ordinary moral beliefs will result in (1) their whole-
sale rejection and (2) the adoption of some mysterious ‘indubitable
starting point’, rather than a set of revised moral beliefs subject to reflec-
tive equilibrium’s usual ongoing epistemic tests. Both of these assump-
tions are quite gratuitous. A more likely result of considering Singer’s
background theory is a reduction in the weight we are willing to assign to
judgments that have a vividly personal quality, such as judgments favour-
ing family members or judgments assigning special opprobrium to harms
inflicted in a direct physical way, as these are likely to have evolutionary
origins (Tersman 2008: 397–8). There may be a corresponding increase
in the weight we are willing to assign to universal or impartial judgments,

4
Singer later seems to make this concession; see de Lazari-Radek and Singer (2012: 29–31).
56 Chapter 4

which have less (or possibly no) evolutionary baggage. Reforming our
judgments in this way would not mean that ‘the “data” that a sound moral
theory is supposed to match have become so changeable that they can
play, at best, a minor role in determining the ﬁnal shape of the normative
moral theory’ (Singer 2005: 349). On the contrary, shifting judgments
play a full role as part of a ‘dynamic dialectical process’ (Brink 2014: 688).

5 How to Use the Method of Reﬂective Equilibrium

In this section I suggest some steps in the process of reflective
equilibrium.
The first step in reaching equilibrium is making considered
judgments on the topic at hand. These are what I take to be the
requirements for considered judgments (Rawls 1999: 42):
1) No upset, fright, tiredness, or intoxication. This may seem obvious, but
there are plenty of cases where political theorists do their work when
subject to personal distress, or to a deadline, or late at night, or (so
I hear) over a glass of wine or two.
2) No conflicts of interest. Individual political theorists often would gain
more under one set of principles than another. Even though there is
effectively no chance that the principles are going to be put into effect,
there may still be a psychological effect. This is actually a rather hard
problem to get around – surely we cannot prohibit work on social
justice, on the basis that any principle would be likely to have effects
on theorists’ incomes. Perhaps the best we can do is be aware of our
possible subconscious biases, and exercise particular caution when
rejecting principles that do not serve our self-interest.
3) The ability to reach the correct decision. This requires at least minimal
standards of competency. It would be possible to reach a reflective
equilibrium about a topic within political theory that one had never
read anything about, but it is unlikely to have much epistemic value
(see Scanlon 2014: 82).5

5
It might be objected that this requirement seems incompatible with my rejection of the
‘epistemic constraint’ in Section 2. This objection misses the importance of the distinction
between constraints on the contents of judgments (such as the epistemic constraint) and
constraints on the circumstances of judgments. The former type of constraint is otiose, as
what it attempts to do (for instance, justification) is done more thoroughly by wide
reflective equilibrium. The latter type of constraint is essential as its functions cannot be
replicated by wide reflective equilibrium proper. For instance, a logical impossibility
should be cleared from our judgments once we consider relevant background theory,
provided we are reasoning in favourable circumstances. But the effects of unfavourable
circumstances, such as being drunk or ignorant of relevant political theory, will not be
cleared by reflective equilibrium, as the epistemic value of the process is fatally undercut
Reflective Equilibrium 57

4) The desire to reach the correct decision. The individual must be motivated
to arrive at justified principles. People (almost?) invariably come to
political theory with a set of preconceived ideas about politics. This is
fine provided that the individual is open-minded, willing to alter his or
her views in response to arguments. The fact that one is on record
defending a position should be no barrier to rejecting that position,
even where this might prove inconvenient or embarrassing.6
In short, the first step is to make sure that you undertake the process of
reflective equilibrium in the Rawlsian ‘conditions favorable for delibera-
tion and judgment in general’. The conditions established in the first
step must be maintained throughout the process.
The second step is to draw up a list of the main contending
principles on whatever topic you are considering. If you can think
of any compelling new principles, these should also be added to the list.
There is no specific number of principles that you should aim for, but as
a general rule and time permitting, more is better. Remember that, while
it is usually impractical in a work of political theory of 5,000 or 10,000
words to discuss a large number of principles, there is no ‘word limit’
when it comes to considering principles prior to or during the actual
writing process. Even if you only discuss two or three principles in detail
in the final product, you may have considered and rejected many more
during the process of reaching equilibrium. Presumably Rawls himself
did – in the ‘Presentation of Alternatives’ section of A Theory of Justice he
names more than a dozen ‘conceptions of justice’, several of them con-
taining multiple principles and one of them the extremely open-ended
‘list of prima facie principles (as appropriate)’ (Rawls 1999: 107). While it
would not be usual to provide such a lengthy list in writing, it is often
useful to mention in passing your reasons for rejecting some of the
principles that do not receive full discussion.
The third step is to begin reflective equilibrium in earnest. You
go through each principle, checking its prescriptions against your
judgments. Ask yourself: what are the central cases for my topic?
And what are the hard cases for this principle? The literature is, of
course, an invaluable resource for finding such cases, but you will also
come up with your own. Consider whether you can accept the impli-
cations of the principle in each of these cases. It may be that initially

by our adverse physical condition or inability to draw on relevant arguments, principles

and theories.
6
A fifth step would be to expose oneself to a wide and representative range of non-
philosophical experiences, in order to offset formative biases. While I am attracted to
this proposal, it does go beyond the method of reflective equilibrium as usually conceived;
DePaul (1993) treats it as part of the separate ‘method of balance of refinement’.
58 Chapter 4

the principle seems to have an unacceptable implication, but that on

reflection you are willing to revise your judgment. This may particularly
be the case where the principle is compelling in other cases. It may seem
to explain why we think what we do in those cases, and extend in an
appealing way to further, previously unconsidered cases. If you can accept
a principle’s implications, either right away or on reflection, then it would
seem that this principle is worthy of further consideration. If you cannot,
you may set aside the principle for now, taking a note of the specific
problems it faces. Repeat this procedure for each principle.
The fourth step is to bring in devices of representation and back-
ground theories (for example, the original position and a theory of the
separateness of persons, respectively). The most important devices of
representation and background theories relevant to the topic should be
considered, with particular devices and theories chosen on the basis of our
judgments, which may themselves be revised during the process, even in
response to normative principles. It may be that you can find a device of
representation that seems, at least on reflection, to capture reasonable
constraints on theory selection. It may even be that you have more con-
fidence in it than you have in any principle. For instance, I personally find
the difference principle less plausible than the original position from which
Rawls controversially (Hare 1975: 102–7; Harsanyi 1975) derives it.
In such cases, you may decide to focus on the principles chosen from the
circumstances specified by the judgment-endorsed device, though you are
still free to directly check the chosen principles against considered judg-
ments (Mandle 2009: 40). Background theories have a similar, though less
dramatic role, guiding principle selection but not outright replacing direct
reference to judgments. Plausible background theories are used at this
point to assess principles, with a particular focus on the principles found
appealing in stage three. Devices of representation are also tested by back-
ground theories. For instance, if we accepted Sandel’s (1982: ch. 1) claim
that the original position assumes that the self is prior to values, we might
reject that device of representation as incompatible with our favoured non-
moral background theory even if it were compatible with our normative
judgments (Gaus 1996: 105).
Having considered principles, devices of representation and back-
ground theories, the fifth step is to review this process. Now you
know the specific challenges faced by the various principles, are
there any revisions to these principles worth considering? Or do
any entirely new principles now come to mind? If so, the third and
fourth steps should be repeated for these principles. If new or revised
principles keep arising, many iterations of the third and fourth steps may
be necessary. The same repetition applies where revised or new devices of
Reflective Equilibrium 59

representation and background theories arise. Likewise, if you have not

found any principles that you find acceptable in their implications, the
third and fourth steps should be repeated. It may, however, be acceptable
to limit the level of repetition due to time constraints. We do not all have
months or years of philosophical contemplation available to us! If all the
steps are followed, the method will yield dividends even if the fifth step is
attenuated, as may be necessary if writing a student essay, for example.
The sixth step is to establish priority rules. This applies only where
you have accepted multiple principles that may come into conflict with
each other. Where you have such principles you need to consider cases of
conflict, and decide how much importance each principle has in it. It may
be that one principle seems so important that it should have absolute or
‘lexical’ priority over another. Alternatively, the principles may seem to
have similar importance, in which case some kind of weighting should be
decided. It could even be found that there are ‘incompatible but equally
justified overall accounts of the subject, thus supporting a kind of plural-
ism about the subject’ (Scanlon 2014: 78–9).
The seventh step is the conclusion of the process, insofar as it
has one. By this point you should have found agreement between prin-
ciples and judgments – or otherwise concluded that this is impossible as
there are no acceptable principles! Either way, your findings are only
ever provisional, and should be considered permanently open to
revision.

6 An Example
I will now work through an example of reflective equilibrium on the topic
of distributive justice, using the aforementioned step-by-step guide and
my own considered judgments. I can obviously give only the scantest
indication of my reasoning here, summarizing years of work in a few
paragraphs. It is likely, furthermore, that the reader will disagree with
me at numerous points. The example should nevertheless illustrate one
way of reaching reflective equilibrium.
The first step is to make sure that my judgments are considered.
As I write this, it is 9.32 am, I had a good night’s sleep, I am aware of
the danger of conflicts of interest when discussing the societal allocation
of goods and am willing to counteract any resulting bias, and I have the
motivation and desire to reach the correct decision. So it seems that, right
now, I am making my judgments in suitable conditions. But as this test
must be taken each time you use the method of reflective equilibrium, it
must be repeated many times – indeed, many thousands of times in my
case!
60 Chapter 4

For the second step I have to draw up a list of the main principles within
this topic. Here’s my list:
– The principle of utility
– Rawls’s two principles of justice
– Equality of outcome
– Luck egalitarian principles (Arneson 1989; Cohen 1989)
– Democratic egalitarian principles (Anderson 1999)
– The principle of priority (Parfit 2000)
– The principle of sufficiency (Frankfurt 1987)
– Right libertarian principles (Nozick 1974: ch. 7)
– Left libertarian principles (Steiner 1994)
– The benefiting principle (Butt 2007)
– The principle of need
– The principle of desert
– Communitarian principles (Sandel 1982)
– Contractarian principles (Gauthier 1986)
– Egoist principles
The list is eclectic, and by design – the point at this stage is to avoid
missing anything important, not to construct the most elegant inventory
possible. Even so, other people’s lists would no doubt contain additional
principles.
With the third step I begin the reflective process by testing the princi-
ples and judgments against each other. Many principles can be set aside
quite quickly. I find nothing of merit in ‘free for all’ egoist principles, and
view the results of contractarian principles for people with low bargaining
power as utterly unacceptable, for instance. I do, by contrast, feel the pull
of the principle of sufficiency, as I am very concerned by those who are
very badly off in absolute terms. But I do not accept its implication that
those who are just below the threshold of ‘having enough’ get absolute
priority over those marginally above the threshold, who are only slightly
better off (Arneson 2006: 28). I therefore conclude that the principle of
priority better accommodates my concern with the absolutely badly off.
Similarly, I am attracted to equality of outcome and democratic equality,
as I am also concerned about inequality. But I am unhappy with demo-
cratic equality’s implication that large unchosen inequalities do not mat-
ter as long as individuals have equal social standing, and equality of
outcome’s implication that, where some squander their equal share of
resources, for instance by deliberately developing ‘expensive tastes’
(Dworkin 2000: 48–59), they should be ‘compensated’ to restore equal-
ity, at society’s expense. I find that luck egalitarianism, which avoids such
problems, fits with my judgments here better, but it has its own appar-
ently objectionable implication that those who make bad choices that
Reflective Equilibrium 61

leave them in severe disadvantage will be ‘abandoned’ (Anderson 1999:

295–6). Outcome egalitarianism has no such implication. So at the end of
the third stage I have a provisional endorsement of prioritarianism, and an
interest in egalitarianism that I am not yet convinced is well expressed in
any principle.
The fourth step sees the introduction of background theories (I set
aside devices of representation). I will mention only one line of thought
here, to illustrate how background theories might help us arbitrate
between political principles. Some critics of luck egalitarianism have
claimed (1) that it assumes that metaphysical libertarianism (the theory
that free, non-causally determined human action is possible) is true, and
(2) that metaphysical libertarianism is false (Scheffler 2003: 17–19).
Were this true, I would have a background theory-based reason to reject
luck egalitarianism in favour of outcome egalitarianism or democratic
equality. However, on reflection I find reasons for rejecting both claims.
While (2) is possible, we do not have adequate grounds for assuming this
to be the case; political theorists would do better to proceed under the
assumption that any of the main theories of free will (including sceptical
views such as hard determinism) might be correct (call this the ‘thin
theory’). Regarding (1), the standard Arneson-Cohen construal of luck
egalitarianism does not after all assume any theory of free will, but is
instead responsive to the morals and metaphysics of responsibility, in the
sense that what counts as ‘chosen’ (and therefore as potential justification
for inequality) depends on the best philosophical account. If metaphysical
libertarianism is false, this just means that one way in which choice might
arise cannot actually happen. Luck egalitarianism would even be compa-
tible with there being no way for true choice to arise. In that case, no
inequality would be justified, a point that mitigates the ‘abandonment
objection’ to luck egalitarianism mentioned in the previous paragraph
(Knight 2015: 132–4). So luck egalitarianism is in fact admirably respon-
sive to what I take as the most plausible background theory about free will,
which is the thin theory (Knight 2009: ch. 5). Outcome egalitarianism
and democratic equality are not responsive in this way, however, as they
make the same prescriptions whether metaphysical libertarianism is true
or hard determinism is true. This seems a significant flaw to me as, in my
judgment, where a person has prima facie brought some hardship upon
herself, we have more reason to assist her if her action were not a true
exercise of free will, and less reason to assist her if her action were a true
exercise of free will. As luck egalitarianism seems to accommodate the
most plausible background theory better than rival egalitarian theories,
I accept it as part of my wide reflective equilibrium.
62 Chapter 4

For simplicity, I leave aside the fifth step. This brings us to the sixth
step. I have found luck egalitarianism and prioritarianism to be in accord
with my judgments. Now we need to decide on a rule to regulate conflicts
between these principles. It seems that neither the ‘eliminate involuntary
disadvantage’ (Cohen 1989: 916) goal of luck egalitarianism, nor prior-
itarianism’s concern with increasing absolute advantage (in particular,
that of the worst off), should be assigned lexical priority, as I would be
willing to give up a small improvement in either of these dimensions for
a large improvement in the other. So my conception of justice is a version
of ‘responsibility-catering prioritarianism’ (Arneson 1999; see also
Knight 2009: ch. 6), where luck egalitarianism and prioritarianism are
balanced against each other. Exactly what weighting, however, to give
each of these principles is a rather tricky question, to be tested through
considering a large number of cases, which I cannot do here.
Suppose, though, that I find a favoured weighting, and reach the
seventh and final step. Even then I cannot assume that weighting, or
even the selection of principles, to be settled for all time, as we can
reconsider any aspect of the process at any point. As Rawls (2005: 97)
cautions, ‘[t]he struggle for reflective equilibrium continues indefinitely.’

References
Ackerman, Bruce, 1980. Social Justice in the Liberal State. New Haven, CT: Yale
University Press.
Anderson, Elizabeth, 1999. ‘What is the point of equality?’, Ethics 109, 287–337.
Arneson, Richard, 1989. ‘Equality and equal opportunity for welfare’, Philosophical
Studies 56, 77–93.
Arneson, Richard, 1999. ‘Equality of opportunity for welfare defended and
recanted’, Journal of Political Philosophy, 7, 488–97.
Arneson, Richard, 2006. ‘Distributive justice and basic capability equality’, in
Alexander Kaufman, ed., Capabilities Equality. Abingdon: Routledge, 17–43.
Brandt, Richard, 1979. A Theory of the Good and the Right. Oxford University Press.
Brink, David, 1989. Moral Realism and the Foundations of Ethics. Cambridge
University Press.
Brink, David, 2014. ‘Principles and intuitions in ethics’, Ethics 124, 665–95.
Butt, Daniel, 2007. ‘On benefiting from injustice’, Canadian Journal of Philosophy
37, 129–52.
Cohen, G. A., 1989. ‘On the currency of egalitarian justice’, Ethics 99, 906–44.
Cummins, Robert, 1998. ‘Reflection on reflective equilibrium’, in Michael DePaul
and William Ramsey, eds., Rethinking Intuition. Lanham, MD: Rowman and
Littlefield, 113–28.
Daniels, Norman, 1996. Justice and Justification. Cambridge University Press.
DePaul, Michael, 1986. ‘Reflective equilibrium and foundationalism’, American
Philosophical Quarterly 23, 59–69.
Reflective Equilibrium 63

DePaul, Michael, 1993. Balance and Reﬁnement: Beyond Coherence Methods of

Moral Inquiry. London: Routledge.
Dworkin, Ronald, 2000. Sovereign Virtue: The Theory and Practice of Equality.
Cambridge, MA: Harvard University Press.
Ebertz, Roger, 1993. ‘Is reflective equilibrium a coherentist model?’, Canadian
Journal of Philosophy 23, 193–214.
Frankfurt, Harry, 1987. ‘Equality as a moral ideal’, Ethics 98, 21–43.
Freeman, Samuel, 2007. Rawls. Abingdon: Routledge.
Gaus, Gerald, 1996. Justificatory Liberalism. Oxford University Press.
Gauthier, David, 1986. Morals by Agreement. Oxford University Press.
Goodman, Nelson, 1965. Fact, Fiction, and Forecast. Indianapolis, IN: Bobbs-
Merrill.
Hare, R. M., 1975. ‘Rawls’ theory of justice’, in Norman Daniels, ed., Reading
Rawls. Oxford: Blackwell, 81–107.
Hare, R. M., 1981. Moral Thinking: Its Levels, Method, and Point. Oxford
University Press.
Harman, Gilbert, 2003. ‘Three trends in moral and political philosophy’, Journal of
Value Inquiry 37, 415–25.
Harsanyi, John, 1975. ‘Can the maximin principle serve as a basis for morality?’,
American Political Science Review 69, 594–606.
Hooker, Brad, 2012. ‘Theory vs anti-theory in ethics’, in Ulrike Heuer and
Gerald Lang, eds., Luck, Value, and Commitment: Themes from the Moral
Philosophy of Bernard Williams. Oxford University Press, 19–40.
Kelly, Thomas and McGrath, Sarah, 2010. ‘Is reflective equilibrium enough?’,
Philosophical Perspectives 24, 325–59.
Knight, Carl, 2006. ‘The method of reflective equilibrium’, Philosophical Papers 35,
209–25.
Knight, Carl, 2009. Luck Egalitarianism: Equality, Responsibility, and Justice.
Edinburgh University Press.
Knight, Carl, 2015. ‘Abandoning the abandonment objection: luck egalitarian
arguments for public insurance’, Res Publica 21, 119–35.
de Lazari-Radek, Katarzyna and Singer, Peter, 2012. ‘The objectivity of ethics and
the unity of practical reason’, Ethics 123, 9–31.
Lewis, David, 1983. Philosophical Papers, volume 1. Oxford University Press.
Maffetone, Sebastiano, 2010. Rawls: An Introduction. Cambridge: Polity.
Mandle, Jon, 2009. Rawls’s A Theory of Justice: An Introduction. Cambridge
University Press.
McDermott, Daniel, 2008. ‘Analytical political philosophy’, in David Leopold and
Marc Stears, eds., Political Theory: Methods and Approaches. Oxford University
Press, 11–28.
Nozick, Robert, 1974. Anarchy, State and Utopia. Oxford: Blackwell.
Parfit, David, 2000. ‘Equality or priority?’, in Matthew Clayton and
Andrew Williams, eds., The Ideal of Equality. Basingstoke: Palgrave, 81–125.
Rawls, John, 1975. ‘The independence of moral theory’, Proceedings and Addresses
of the American Philosophical Association 48, 5–22.
Rawls, John, 1999. A Theory of Justice, revised edition. Oxford University Press.
64 Chapter 4

Rawls, John, 2005. Political Liberalism, expanded edition. New York: Columbia
University Press.
Sandberg, Joakim and Juth, Niklas., 2011. ‘Ethics and intuitions: a reply to Singer’,
Journal of Ethics 15, 209–26.
Sandel, Michael, 1982. Liberalism and the Limits of Justice. Cambridge University
Press.
Scanlon, T. M., 2003. ‘Rawls on justification’, in Samuel Freeman, ed.,
The Cambridge Companion to Rawls. Cambridge University Press, 139–67.
Scanlon, T. M., 2014. Being Realistic about Reasons. Oxford University Press.
Scheffler, Samuel, 2003. ‘What is egalitarianism?’, Philosophy and Public Affairs
31, 5–39.
Singer, Peter, 1974. ‘Sidgwick and reflective equilibrium’, The Monist 58,
490–517.
Singer, Peter, 2005. ‘Ethics and intuitions’, Journal of Ethics 9, 331–52.
Sinnott-Armstrong, Walter, Young, Liane and Cushman, Fiery, 2010. ‘Moral
intuitions’, in John Doris and The Moral Psychology Research Group, eds.,
The Moral Psychology Handbook. Oxford University Press, 246–72.
Steiner, Hillel, 1994. An Essay on Rights. Oxford: Blackwell.
Tersman, Folke, 2008. ‘The reliability of moral intuitions’, Australasian Journal of
Philosophy 86, 389–405.
Varner, Gary, 2012. Personhood, Ethics, and Animal Cognition. Oxford University
Press.
Walden, Kenneth, 2013. ‘In defense of reflective equilibrium’, Philosophical Studies
166, 243–56.
Welch, John, 2014. Moral Strata: Another Approach to Reflective Equilibrium.
Dordrecht: Springer.
Williamson, Timothy, 2007. The Philosophy of Philosophy. Oxford: Blackwell.
Wolff, Jonathan, 2013. ‘Analytic political philosophy’, in Michael Beaney, ed.,
The Oxford Handbook of the History of Analytic Philosophy. Oxford University
Press, 795–822.
5 Contractualism

Jonathan Quong

1 Introduction
Are you a contractualist? People sometimes ask this question in the same
way you might ask someone, are you Muslim or are you a Utilitarian?
They ask as if contractualism is a deep substantive commitment – as if it’s
an exclusive doctrine that competes with other normative theories.
I believe that this way of thinking about contractualism is flawed.
Contractualism is a method or a tool, not a first-order moral theory.
Asking a philosopher if she is a contractualist is thus a bit like asking
a carpenter if she is a ‘Hammerist’, that is, someone who uses a hammer to
the exclusion of all other tools of the trade.
For the purposes of this chapter, I will assume that contractualism is
a method used to answer some first-order normative question in moral or
political philosophy, according to which we should consider what a group
of idealized agents would or could accept under certain specified con-
straints. I will refer to the idealized agents as the contractors, and I will refer
to the general choice situation in which they are placed as the contractual
scenario. Obviously many different forms of contractualism are possible
since different assumptions can be made regarding who the contractors
should represent, how they should be idealized, what question they
should be tasked to answer, what constraints they should face, and what
alternatives they should confront in the event of non-agreement.
In this chapter I defend three main claims. First, like any good tool or
method, contractualism is a helpful way to tackle some problems, but not
others. Second, there are very different types of contractualism, and these
different types are designed to address different normative questions.
Finally, and relatedly, work in moral and political philosophy sometimes
goes wrong due to the failure to carefully distinguish these different forms
of contractualism.
The remainder of this chapter is organized as follows. Section 2 argues
that we can usefully distinguish three different forms of contractualism:
consent contractualism, fairness contractualism, and rationality contractualism.

65
66 Chapter 5

Section 3 considers how contractual scenarios should be designed in light of

the particular question we’re trying to address. Section 4 identiﬁes some
ways contractualism can be misused. Section 5 concludes by summarizing
this chapter’s central claims.

2 Why Contractualism?
Why ask what suitably situated contractors would or could accept? Why
not just ask what real people do in fact accept?
It’s not always possible to ask the relevant people what they would
accept. Sometimes this is because we are deciding whether to surprise
someone, whether to reveal information to someone, or whether to lie
to someone. Seeking the person’s consent would make it impossible to
surprise the person, avoid giving the person the information, or lie to
the person. Other times the problem is that the relevant people don’t
currently exist: we cannot directly talk to past or future generations.
Even when we can communicate with the relevant people, we may have
good reasons to disregard the answers they will provide. If the question
that interests us is what P would agree to if she were perfectly rational,
then asking the real P, who is far from perfectly rational, is not a good way
to answer the question. Similarly, we might want to know what political
principle persons would accept if everyone had exactly equal bargaining
power. Because real people do not have exactly equal bargaining power,
we cannot pose our question directly. We might also have good reasons to
ask what persons would accept if they had somewhat different aims or
motives – for example, what moral rules people would accept if they did
not already have various ﬁrm beliefs about rightness and wrongness, or
what political rules people would accept if they all shared a few particular
ideals, but diverged in many other ways.
In sum, it may be either impossible or undesirable to ask what real
people would accept. Under these conditions contractualism may be an
appropriate method for tackling certain moral or political questions. But
when exactly is it an appropriate method for doing so? What questions is
contractualism well suited to help us answer?
To begin, contractualism can be a useful way of trying to understand
whether some presumptively wrongful act or rule, ϕ, is in fact permissible
or legitimate when we think that permissibility or legitimacy depends on
whether persons would freely consent to ϕ if they could. For example, it is
presumptively wrong to cut someone open and operate on him without
his consent, but if the person freely consents to the surgery, the consent
typically renders the surgery permissible. But sometimes we cannot
obtain the person’s consent, and yet there are powerful reasons to proceed
Contractualism 67

with the surgery: an unconscious man on the road needs immediate

medical attention, but obtaining his actual consent isn’t possible.
In cases like this, it might seem that we should ask whether we believe
this person would consent to the emergency treatment, and if the answer is
yes, this is sufﬁcient to render the surgery permissible.
There are also situations involving larger numbers of people that may
have a similar structure. For example, it is presumptively wrong to coerce
others to obey commands (at least where the commands are not indepen-
dent requirements of justice), unless those subject to the coercion consent
to the relevant authority. But this is how the state behaves: it coerces
citizens to obey a large number of commands, many of which are not
independent requirements of justice, and yet most citizens do not consent
to the authority of their state, nor do most citizens have a realistic oppor-
tunity to freely give or withhold their consent. Some have thus suggested
that the way to vindicate at least some of the modern state’s coercive
authority is to consider whether suitably situated contractors would con-
sent to its authority.
These cases share the following structure. It is presumptively wrong to
do something, ϕ, to a person or group of persons unless those people
freely consent to it. But in circumstances C, there seem to be good reasons
to ϕ, and yet obtaining the actual consent of the relevant people is not
possible. When faced with a problem of this kind, asking whether suitably
situated contractors would consent might be a helpful way of deciding
whether it is permissible to ϕ. I will call the version of contractualism that
is designed to address this problem consent contractualism.
Contractualism might also be invoked, however, for other reasons.
Suppose, for example, we are playing a low-stakes game of poker, and
in the middle of the game, we realize that the seven of clubs has been
missing from the deck the entire time (Dworkin 1973: 501). A question
now arises: what should be done? Should we treat the results of the game
thus far as legitimate, or should we declare the results of the game so far to
be void and begin again, with each player being returned to his or her
initial allocation of chips? Asking the current players what they think is
fairest might not be a reliable way of answering this question, since their
answers are likely to be inﬂuenced by their self-interest, by how many
chips they have already won or lost. Contractualism seems like a natural
way to approach the problem. To decide what the fairest rule would be,
we ask what suitably situated contractors – for instance, contractors who
do not have a self-interested reason to prefer a particular outcome – would
all accept.
Of course, everyone in philosophy is familiar with the most famous
version of this model: Rawls’s use of the original position and the veil of
68 Chapter 5

ignorance as a method for determining the best principles of justice to

regulate the distribution of the burdens and benefits of social cooperation
amongst citizens (Rawls 1999: 102–68). We can call this fairness contrac-
tualism. We can also envisage a more general version of fairness contrac-
tualism, where the contractual device is used to determine what would be
morally right or permissible in a given context, and not merely what
would be fair (Scanlon 1998). Although such versions of contractualism
are concerned with more than fairness, they can without too much dis-
tortion be included as variants of fairness contractualism since they share
the same basic structure: the contractual device is designed to reflect
some of our existing moral judgements in order to help us answer further
moral questions about what is fair or right.
Although fairness contractualism and consent contractualism are often
conflated, they differ in several important respects. First, the aim of fair-
ness contractualism is not to determine whether some presumptively
wrongful act or rule is morally permissible. The aim is rather to determine
what fairness requires in a given context. A rule may be fair, and yet it may
be impermissible to impose the rule without the actual consent of the
persons involved. Conversely, it may be permissible to perform acts or
impose rules on others even when doing so does not meet some standard
of fairness. Fairness is thus neither necessary nor sufficient to settle
questions about permissibility. Second, unlike consent contractualism,
the success of fairness contractualism does not depend on how closely the
contractors approximate the real persons they represent. The aim of
fairness contractualism is not to establish, as far as is possible, whether
some real people would consent to a given rule or policy. The aim is rather
to model a contractual scenario such that it is guaranteed to produce a fair
agreement. The contractors may thus depart radically from their real-
world counterparts, and so it may often be true that we cannot infer
anything about the consent or the will of real persons from the consent
of contractors. But this is no objection to fairness contractualism, since
the aim is not to accurately model or represent the consent or will of
particular real persons, but rather to model a perfectly fair contractual
scenario.
Sometimes, however, the question that interests us is neither the con-
sent of real people, nor what people would accept under perfectly fair
conditions. Instead, sometimes we want to understand what real people
would all accept if they were perfectly rational, either in the narrow
instrumental sense of choosing the most effective means to pursue their
self-interest, or in the broader sense of making the most rational choice in
light of all their beliefs and commitments. Suppose, for example, there’s
an isolated town with just one large employer: OmniCorp. Everyone who
Contractualism 69

works for OmniCorp is required to contribute to a compulsory insurance

scheme that provides a certain level of medical coverage in the event of
serious illness or injury. We might ask many different questions about this
compulsory insurance scheme, but one question is whether the scheme is
one that all the employees would accept if they were self-interested and
instrumentally rational. If the scheme cannot be defended on the basis
that it is in everyone’s self-interest, then it will have to be defended on
other grounds. But if the scheme is in everyone’s self-interest, then
(depending on our broader moral theory) it may be irrelevant that the
people in the town lack alternative employment options, and so have no
realistic choice but to ‘consent’ to OmniCorp’s insurance scheme.
This example is, in certain key respects, a small-scale version of the
political situation in which many citizens find themselves. Many citizens
have no realistic opportunity to refuse the package of burdens and ben-
efits the state offers, and so they cannot meaningfully consent to it. But we
can ask whether the package the state offers is prudentially rational for
citizens to accept. For example, can some citizens plausibly complain
that, were it possible, they would have been able to bargain for better
terms prior to leaving the state of nature and entering civil society? Some
political philosophers, most notably Thomas Hobbes and David
Gauthier, focus on questions of this kind – on determining whether the
moral or political rules that govern our lives can be partly or entirely
vindicated by establishing that they are terms that would be rational for
each person to accept even if each was self-interested (Gauthier 1987;
Hobbes 1996). A closely related, but broader, question pursued by some
contractualists is this: what moral or political rules would each person
have sufficient reason to accept in light of all her beliefs and commitments
(Gaus 2011)? We can call all forms of contractualism designed to con-
sider what people would accept if they were rational, either in a narrow
self-interested sense, or in some broader sense, rationality contractualism.1
Rationality contractualism differs from consent contractualism in at
least two important respects. First, the aim is not to use a contractual
device to ascertain whether real people would in fact consent to some rule
or policy. Rather, the aim is to ascertain whether they rationally ought to
accept some rule or policy given certain aims. In focusing on what would
be rational, this form of contractualism deliberately sets aside various
things – allegedly independent moral commitments, ignorance, and irra-
tionality – that might influence real people’s decisions. Second, ration-
ality contractualism need not focus on acts or rules that are presumptively

1
Some refer to the narrow self-interested form of rationality contractualism as contractar-
ianism and contrast it with the contractualism of Rawls and Scanlon.
70 Chapter 5

wrongful absent individual consent. Rationality contractualism can take

as its focus any act, policy, or set of institutions.
Rationality contractualism also clearly differs from fairness contractu-
alism. Unlike fairness contractualism, rationality contractualism is not an
effort to model our ideas about fairness. What people would accept under
perfectly fair conditions may depart dramatically from what it would be
rational for them to accept given their actual circumstances and given
some particular set of aims.

3 Designing Contractualism
Having identiﬁed three of the main reasons to use contractualism, we can
consider in greater detail how contractualist scenarios should be designed
to best address a given problem. It will be helpful to distinguish ﬁve key
questions.
i. Constituency: Who do the contractors represent?
ii. Primary Question: What question do the contractors confront?
iii. Motives and Interests: What are the motives and interests of the
contractors?
iv. Information: How much information should be provided to the
contractors?
v. Non-agreement: What will happen to each contractor in the event of
non-agreement?
These are not the only questions that matter in designing a model of
contractualism, but they are central questions that heavily shape how the
model will work and what it can achieve. Let’s now consider how each of
these questions might be addressed from the perspective of the different
forms of contractualism.

3.1 Consent Contractualism

Consent contractualism aims to address cases that have the following
structure: it is presumptively wrong to do something, ϕ, to a person or
group of persons unless those people freely consent, but in circumstances
C, there seem to be powerful reasons to ϕ, and yet obtaining the actual
consent of the relevant people in C is not possible.
The constituency of consent contractualism is thus usually not difficult
to determine: every real person who would have a presumptive veto
against ϕ must be represented in the contractual scenario. Suppose, for
example, the question is this: may the state take action now to avert
a significant environmental disaster, if the consequence of doing so is
that a particular group of people who are currently five years old or
Contractualism 71

younger will suffer a greatly increased chance of being killed prematurely

by radiation poisoning forty years from now?2 If you believe it is pre-
sumptively wrong to take action that will greatly increase a person’s
chances of being killed prematurely by radiation poisoning, then consent
contractualism might be well situated to address this case, since the
people whose consent is of normative signiﬁcance are currently too
young to be consulted.
Next, let’s consider the primary question contractors confront.
As theorists, we want to know if imposing ϕ is morally permissible, or at
least whether doing so would infringe the rights of those on whom ϕ is
imposed. But this isn’t the question the contractors are tasked to answer.
They are not asked: is imposing ϕ permissible or an infringement of your
rights? We don’t pose this question to the contractors since the contrac-
tors would then need to have in hand the relevant moral answer that we’re
seeking to ﬁnd via the contractual device. Instead we ask the contractors
a different question: would you agree to the imposition of ϕ? If all the
contractors can agree, then we might infer that the real people they
represent would consent if they were in a position to do so. Of course,
whether the contractors agree to the imposition of ϕ will depend on
further features of the contractual scenario.
In consent contractualism, it’s typically important that the motives and
interests of the contractors mirror, as closely as possible, the motives and
interests of the real people the contractors are meant to represent. In these
cases, consent contractualism is used as a proxy for real consent, but this
won’t be plausible if the motives and interests of contractors depart too
dramatically from those of the real persons they represent.
There are, however, several worries about this strategy. First, some-
times it’s not possible to know much about the motives and interests of
the real people in question. But this need not be an insurmountable
problem. We can often make plausible general assumptions about the
interests and motives of people about whom we don’t otherwise know
very much. We can assume, for example, that people strongly prefer not
to die prematurely, and that, other things being equal, they prefer to have
a greater range of legal liberties. If the consent of the parties turns only on
these sorts of general interests – ones that can be safely assumed – then it
may not matter that we lack detailed information about the motives and
interests of the real people.

2
In this example the identity of future persons is not what is at issue. Whether contrac-
tualism can also be used to address the so-called non-identity problem (where the choice
of policy will affect the identity of the future persons) is a further question. For an example
where contractualism is used in this way, see Reiman (2007).
72 Chapter 5

A different worry, however, is that the real people whose consent

matters may have morally repugnant motives or interests. Shouldn’t we
prevent contractors from having repugnant aims or interests to ensure the
contractual device does not yield clearly unjust or immoral results? For
example, it is presumptively wrong to gently pinch a person, P, on the
arm, but many things can render this act permissible even if P does not
consent. Suppose doing so is the only way to save twenty children from
being burned alive. It’s circumstances like this, I take it, that motivate the
worry about excluding certain wrongful motives or interests from the
contractual scenario: we don’t want a contractor to be able to withhold
consent in this case because she’s spectacularly self-interested or a racist.
But cases like this don’t pose a problem. Under these conditions, P is
morally required to suffer the pinch – she cannot rightfully refuse, and so
her consent (or lack of it) is irrelevant. A case like this is thus not one
where the act is presumptively wrongful without P’s consent. In cases
where the act or rule really is presumptively wrong without the consent of
P, the content of P’s particular reasons for withholding consent are
typically irrelevant. For example, it’s wrong to force P to attend
a religious service for paternalistic reasons unless P consents. P may
withhold consent for morally relevant reasons (e.g. she doesn’t believe
in God) or for morally pernicious reasons (e.g. she is a racist and doesn’t
want to attend a religious service with members of some racial group);
either way, P’s lack of consent is decisive. Provided consent contractual-
ism addresses only those cases where a proposed act or rule is presump-
tively wrongful without the consent of some person or group, then there is
no need to apply moral restrictions to the motives or interests of the
contractors.
We can now turn to a different question: how much information should
the contractors be given? The answer is, as much as is needed to ensure
the contractors’ decision accurately reflects the wills of the real people
they represent. We want to know what real people would accept if they
were confronted with a certain decision, and so contractors should be
given all the information that bears on the decision, where relevance is
determined by appeal to the motives and interests of the persons in
question. If we want to know whether Albert would consent to the
imposition of ϕ, and we believe that Albert’s choice would be influenced
by whether some fact, F, obtains, then the contractor representing Albert
should know whether F obtains when making the choice on Albert’s
behalf. Doing so ensures the contractor’s choice is more likely to accu-
rately reflect Albert’s will.
This, however, is perfectly compatible with using consent contractual-
ism in cases where the relevant decision is one where we don’t have access
Contractualism 73

to all the relevant facts. Suppose, for example, we want to know whether an
unconscious patient would consent to an emergency procedure that has
a 5 per cent risk of death, or whether she would opt against the procedure,
in which case there is a 95 per cent chance she will suffer permanent
paralysis of her upper body. In cases like this it’s clearly acceptable if the
contractors know only the probabilities. After all, this is all the information
we have, and what we want to know is how the real person would choose
given the probabilities, not how she would choose if she knew the outcome.
Finally, we can turn to the issue of non-agreement in consent contrac-
tualism. If contractors are being asked whether they can accept the
imposition of ϕ, clearly a crucial feature of their decision will be informa-
tion regarding what will happen in the event that ϕ is not imposed. Other
things being equal, the worse the non-agreement point is for a given
contractor, the more likely she is to accept the imposition of ϕ. As we
did when considering how much information the contractors should be
given, we should be guided by the aim of ensuring that the choice made by
each contractor reflects the will of the real person being represented as
closely as possible. Thus, as a general rule, the non-agreement point
should be designed to accurately reflect what would happen to someone
in the event that ϕ is not imposed.
In some simple cases, the non-agreement point is clear. For example, in
the case just presented, if the doctor does not proceed with the surgery,
then no surgery is performed, and the patient faces a 95 per cent chance of
paralysis. In other cases, however, determining the non-agreement point
is more difficult. For example, suppose the question being posed is
whether citizens would consent to the coercive authority of the state if
given the choice. What should we assume is the non-agreement point?
Many philosophers have adopted some version of the state of nature as
the non-agreement point, but accounts of the state of nature vary wildly
from Hobbesian to Rousseauian.
The problem runs deeper than the mere fact that philosophers disagree
about what a state of nature might look like. Real citizens in contemporary
liberal democratic societies do not face the stark choice that some philo-
sophers ask their contractors to make: live in a state of nature or under the
authority of a state. Thus, even if we had a clear picture regarding what
the state of nature might look like, we should be wary of inferring anything
about the will or consent of real citizens from the hypothetical consent of
contractors confronting the huge question of whether the state is prefer-
able to the state of nature. Given the aims of consent contractualism, it’s
preferable to focus on questions that can be answered by confronting the
contractors with roughly the same options as those faced by the real
persons they represent.
74 Chapter 5

3.2 Fairness Contractualism

In fairness contractualism, the aim is to situate contractors in such a way
that the choice of a rule or set of rules, R, to govern a situation gives us
good reason to believe that R is fair, just, or morally correct (I will focus on
fairness, for simplicity). Given this aim, let’s consider our ﬁve main
questions.
First, who should be included in the constituency of fairness contractu-
alism? Suppose, for example, we are using fairness contractualism to
determine the fairest principles for regulating the distribution of certain
resources. Who should the contractors represent? One initially tempting
answer is that they should represent everyone who is affected by the rules
that will regulate the distribution of resources. But this may not always be
correct. Suppose you and I work together to achieve some common goal,
and we must decide how to distribute the burdens and beneﬁts between
us. There may be others who are affected by the choice of principles (e.g.
my aunt will get a more generous Christmas present if one set of principles
is selected), but this seems irrelevant to what constitutes a fair distribution
of our collective efforts: my aunt shouldn’t be represented in the contrac-
tual scenario. There is nothing in the logic of contractualism itself that
dictates the constituency of persons who ought to be represented within
a model of fairness contractualism. The appropriate constituency for any
model of fairness contractualism must be determined by independent
ideas regarding who has the standing to be wronged or treated unfairly by
some proposed rules regulating conduct or the distribution of advantages.
Sometimes we may have good reason to believe that everyone affected by
a proposed principle has the standing to be wronged or unfairly treated,
and thus we will want all affected persons to be represented. Other times,
we may have reason to believe that only those who qualify as participants
in some scheme, or only those who are subject to the authority of the
rules, have the requisite standing.
Second, what is the primary question the contractors are asked to
address? As with consent contractualism, we do not pose contractors
the same question we, as theorists, are trying to answer. If we simply ask
contractors what fairness requires, then the contractual device is no
longer doing any helpful work – we would be directly asking ourselves
what fairness requires, and simply putting the answer in the mouth of an
imagined contractor. Instead, we ask contractors what rules regulating
some domain they would all accept if they were either entirely, or at least
partially, self-interested. The further features of the contractual scenario
are then meant to ensure that the resulting agreement captures what
fairness requires. A sceptic may wonder, of course, why we should bother
Contractualism 75

with the contractualist device here. Why not consider the question
directly? I want to set aside this objection for now, but I address it in
Section 4.4.
Let’s move on to the motives and interests of contractors in fairness
contractualism. Contractors must be partly driven by self-interest, at
least in the broad sense described earlier. That is, they must have their
own distinct plans that they wish to pursue. They will thus be motivated
to prefer rules that are more likely to advance their plans. But fairness
contractualism can also allow the motives and interests of contractors to
depart quite radically from the real-world persons that they represent.
The reason is simple: fairness contractualism does not aspire to accurately
reﬂect the will or prudential rationality of real people. Instead the aim is to
design a contractual scenario in such a way that whatever emerges as
a decision is guaranteed to constitute a just or fair decision. One way to
achieve this result is to stipulate that contractors are, in addition to being
self-interested, partly motivated by a desire to justify themselves to others,
and thus will only reject proposals on reasonable grounds, where ‘reason-
able’ is a partly moralized notion, one that requires contractors to take
due account of the interests and status of others affected by a proposed
rule or its rejection (see Scanlon 1998: 189–247).
There are two important and related objections to attributing to the
contractors moral motives that constrain the pursuit of their self-interest.
First, sceptics complain that the appeal to what contractors could reason-
ably reject will leave the results of the contractual device radically inde-
terminate. The problem is that the ordinary or everyday notion of
reasonableness is indeterminate or contested. Of course, there are some
attitudes and behaviours that we can all agree are unreasonable, for
example, viewing oneself as exempt from general moral rules that every-
one else must follow, or behaving as if one’s claims to resources or
advantages carry greater weight than those of similarly situated others.
But the cases in which we agree are too few to yield a precise account of
the concept, such that it can be reliably deployed to evaluate whether the
exercise of a veto is reasonable or unreasonable. Suppose, for example,
the question is whether those who fare worst under some distributive
scheme, D1, can reasonably reject D1 in favour of D2, where they would
be marginally better off, but everyone else would be signiﬁcantly worse off
than under D1, though still better off than the worst off. The ordinary
sense of reasonableness does not provide an obvious answer to this ques-
tion, or to countless other questions where different considerations
appear to pull in different directions. If the notion of the reasonable is
so imprecise, our contractual device will be of no use for reaching deter-
minate results in most cases. This objection need not be limited to
76 Chapter 5

Scanlon’s use of the notion of reasonableness. The objection will apply to

any version of contractualism that attributes a moral motive to the con-
tractors, without clearly specifying how the motive constrains the deci-
sions contractors will make.
One way to defuse this objection is to give the moralized motive very
specific content. We could, for example, specify that what it means to be
reasonable is that one will only veto proposals when there is an alternative
proposal that is anonymously Pareto superior, that is, one where no social
position is worse off, and at least one position is better off. Offering such
a specific account of what reasonableness requires in terms of exercising
one’s veto would defuse the charge of indeterminacy.
But this solution invites another, perhaps even more serious objection,
namely, that the contractual device is now redundant. There’s no longer
any need to consider what contractors might accept to help understand
what fairness requires: we already know what fairness requires – the
anonymous Pareto principle – and we’ve simply told contractors that
they may only veto proposals that violate this principle. Sceptics thus
argue that models of contractualism that impute moral motives to the
contractors face a dilemma: either the stipulated moral motive is too
vague or imprecise to deliver determinate results, or else the motive is
so clearly defined that it renders the contractual device entirely
unnecessary.
As with many alleged dilemmas, one response is a Goldilocks solution:
identify a moral motivation that is sufficiently clear so as to make deter-
minate results possible, without being so precisely defined that it amounts
to giving contractors the exact answer the contractual device is meant to
help us uncover. Whether this happy medium is possible is a crucial
question for anyone who uses a contractual device that attributes some
moral motives to the contractors.
An alternative solution is to restrict the information available to con-
tractors instead of imbuing them with moral motives. Rawls offers the
most influential version of this approach. Contractors are placed behind
a veil of ignorance, and are thus unaware of any specific information
about the plan of life, sex, race, or social or economic position of the
person they represent. By depriving contractors of any specific informa-
tion about the person they represent, each contractor is forced to consider
how proposed rules will affect every possible social position. The resulting
agreement is one that will therefore be fair, since each person’s interests
will have been fully taken into account without allowing any person to
unreasonably veto proposals.
Although this approach may avoid some of the worries raised by imput-
ing moral motives to the contractors, it is not without its own difficulties.
Contractualism 77

I won’t attempt a survey of the many objections that have been pressed
against Rawls’s use of the veil of ignorance. Instead, I want to consider
how much information should be excluded from what the contractors
know when designing a contractual scenario in which contractors are
purely self-interested.
The answer to this question might, at first glance, seem very straight-
forward. We should exclude any information that would give a contractor
an incentive to veto proposals for what we deem morally irrelevant rea-
sons. For example, if contractors knew the race of the person they repre-
sent, they would be inclined to veto proposals that fail to discriminate in
favour of members of their racial group. Since we don’t believe race is
a morally relevant basis for allocating resources, we prevent contractors
from having this information. But things aren’t so simple. Suppose we did
allow each contractor to know the race of the person she represents. This
won’t yield a morally problematic result, where some people are given
a greater share of resources merely on account of race, because any
proposal to benefit a particular racial group will be vetoed by contractors
representing people from other races. The only proposals that look cap-
able of evading a veto are those that disregard race as a relevant consid-
eration for the distribution of resources – any proposal that takes race into
account will be vetoed by at least one contractor. The rationale for
excluding information such as a person’s race thus cannot be that this is
the only way to avoid principles that wrongly allow race to serve as a basis
for the distribution of advantages and disadvantages.
You might protest that allowing contractors to know the race of the
person they represent will still prove problematic because it will lead to
a null set, that is, the contractors will be unable to agree on anything. Even
if racist proposals will be successfully vetoed, won’t non-racist policies
also be vetoed by contractors who will only consent to policies that give
priority to the racial group of the person they represent? Once contractors
have some specific information about which policies would be guaranteed
to maximally benefit the person they represent, won’t they refuse to
consent to anything less?
This problem can be avoided if the non-agreement point for each con-
tractor is less attractive than some of the proposals being considered.
Suppose, for example, contractors are asked to choose amongst potential
principles of distributive justice, and they know that if they cannot agree
on any proposed principles, the non-agreement point is roughly like the
Hobbesian state of nature. Faced with this option, contractors will readily
accept a principle of justice that ignores race as a basis for distribution
even if the contractors know the race of the person they represent and are
purely self-interested. Of course each contractor would prefer a principle
78 Chapter 5

of justice that favours his own race, but each knows the others can veto
any such principle, and each is strongly motivated to reach some agree-
ment to avoid the Hobbesian state of nature. The underappreciated point
is this: imposing a thick Rawlsian veil of ignorance is not necessary to
avoid the sort of partiality and prejudice that is incompatible with fairness,
even when contractors are strictly self-interested. Principles that unfairly
favour a particular race, gender, or religion can be avoided with
a combination of contractors’ vetoes and a non-agreement point that is
equally unappealing for all the contractors.
Some critics have complained that fairness contractualism cannot sim-
ply stipulate a non-agreement point that is sufficiently bad that contrac-
tors will prefer at least one of the proposals being considered. These
critics suggest that genuine contractualism depends on showing that the
agreed principles will be mutually beneficial, where the ‘benefit’ must be
measured according to a benchmark of how well off the real people being
represented could expect to be in the event of non-agreement (Cohen
1995: 225; Nozick 1974: 192–7). We cannot simply stipulate that all
contractors face the same very unappealing non-agreement point. This
stipulation will, in many cases, be false – for example, sometimes some
subset of the real people being represented will do better in the event of
non-agreement than they would under some fair set of distributive prin-
ciples. If so, argue the critics, then the would-be contractualist must
admit her contractualism is a sham in the sense that it fails to deliver
mutually advantageous principles.
But this objection misunderstands the aims of fairness contractualism
(Quong 2007: 77–82). As I’ve already emphasized, fairness contractual-
ism, properly construed, doesn’t aspire to track the wills of real people,
nor does it purport to establish that the principles selected are instrumen-
tally rational for real people given their current aims. The goal is rather to
construct a scenario where self-interested agents must come to an agree-
ment regarding a rule or set of rules, but the scenario is designed in such
a way that the agreement that emerges is guaranteed to be fair. In this
version of contractualism, nothing turns on showing that some real per-
son would benefit from the agreement relative to the status quo or relative
to a realistic non-agreement point.
Let’s return to our earlier observation that depriving the contractors of
all specific information about the persons they represent is inessential to
ensuring that the principles remain untainted by prejudice or wrongful
partiality. If this is true, why does Rawls impose a thick veil of ignorance
on the contractors? I think the answer to this question is, at least in part,
because doing so is a better heuristic for taking up the perspective of
others. Asking ‘what rules would I choose to regulate this domain if
Contractualism 79

I didn’t know which person I would be?’ is a powerful way of forcing

oneself to empathetically take up each person’s perspective. But asking,
‘what rules would a group of self-interested contractors choose to regulate
this domain if they were precluded by each other’s vetoes from favouring
selfish or unduly partial principles, but were still sufficiently motivated to
come to an agreement?’ is less likely to trigger the same taking up of
others’ perspectives. There are also other reasons to impose a thick veil of
ignorance. Doing so is a way of representing what Rawls takes as funda-
mental to the freedom and equality of persons or citizens: a capacity for
a sense of justice and a capacity for a conception of the good. But this,
though important in Rawls’s theory, is not essential to the more general
idea of fairness contractualism.
In sum, there are several ways that a contractual scenario can be
designed to try and ensure that the agreement that emerges is fair: (i)
a moral motive can be directly imputed to the contractors, (ii) the con-
tractors can be deprived of any specific information about the persons
they represent, or (iii) a non-agreement point can be stipulated that is
sufficiently bad such that the contractors will be motivated to arrive at an
agreement and avoid insisting on proposals that others are certain to veto.
Finally, it’s worth emphasizing that while Rawls and Scanlon use the
contractual device to develop theories of justice or interpersonal morality,
fairness contractualism can be used to pursue more modest objectives, for
example, deciding what rules should regulate some specific project
amongst neighbours, or how the costs created by an unintended factory
accident should be distributed amongst the relevant agents.

3.3 Rationality Contractualism

Finally, let’s consider our ﬁve questions from the perspective of ration-
ality contractualism, whose aim is to determine what people would all
accept if they were rational, either in the sense of most effectively pursuing
their self-interest, or else in the broader sense of choosing the best option
in light of all their aims and commitments. Because it has been so
inﬂuential in political philosophy, I will focus mostly on the narrower self-
interested form of rationality contractualism, but we shouldn’t lose sight
of the fact that rationality contractualism needn’t be construed in this
narrower way.
First, what is the constituency of the contractual scenario? The answer to
this question can vary. We may, like Gauthier, be interested in a very
broad question – determining which set of moral rules is instrumentally
rational for everyone to follow – in which case the constituency will
include all (or almost all) persons. But we might be interested in
80 Chapter 5

a much narrower question, for example: is it instrumentally rational for

self-interested agents in a particular neighbourhood to accept the author-
ity of law enforcement officials in their city? Clearly, the constituency of
rationality contractualism can be broad or narrow – it all depends on
whose rationality is of interest to the theorist.
Second, what is the primary question that contractors are given? Again
the answer to this question will vary depending on our more specific
objectives. But the question will usually take one of two forms. Either
we are interested in whether the status quo is rationally acceptable to
persons given what would most likely occur in the absence of the status
quo. Or else the theorist has a more open-ended question in mind, for
example, what rule or set of rules to regulate some domain would people
all agree to if they were instrumentally rational, regardless of what the
status quo happens to be? This more open-ended question, as we’ll see
later, creates a particular challenge regarding the non-agreement point.
Third, in the most influential versions of rationality contractualism, the
motives and interests of the contractors are always self-interested.
To borrow Rawls’s term, the contractors are mutually disinterested:
they take no interest in the plans of others – their goal is simply to advance
their own aims. Can the contractors have, as one of their general aims, to
act in accordance with moral or political rules? If the theorist’s wider
objective is to uncover whether some alleged moral or political rule
conflicts or is congruent with individuals’ rational self-interest, then con-
tractors clearly cannot include conformity with the rule among their aims,
since this makes the answer to the theorist’s question trivial. But this
doesn’t mean that rationality contractualism precludes contractors from
having morally motivated aims. Many people’s central life plans are
shaped by moral convictions: a belief that one ought to act as God
commands, or that one ought to do as much as possible to alleviate
human suffering. These are clearly morally motivated aims, and the
broader forms of rationality contractualism allow contractors to have
such objectives, but within the model, they will be treated as on a par
with any other interests a contractor might have, for example, the self-
interested pursuit of wealth or power.
Apart from these constraints, the aims of the contractors should mirror
those of the persons they represent, though again, how closely the aims of
the contractors should mirror those of the persons they represent will
depend on the theorist’s objective. The theorist may be interested in
taking people roughly as they are – that is, in taking people’s particular
aims or interests as given. The objective is then to uncover how people
with those particular aims would choose if they were instrumentally
rational. But rationality contractualism need not take people exactly as
Contractualism 81

they are. Instead, the theorist might wish to ‘clean up’ people’s existing
aims and interests – correcting for obvious inconsistencies, and discard-
ing those plans individuals have formed that are instrumental to achieve
one objective, but that are not the most effective means of pursuing
a more fundamental objective. There are varying degrees to which the
contractors can be idealized in this way. A moderate form of idealization
might seek to eliminate only obvious inconsistencies and clearly fallacious
inferences made by real persons, while leaving all the person’s central
commitments untouched.3 More radical forms of idealization are possi-
ble, however. In principle, all of a person’s plans or beliefs, apart from her
final ends, could be subject to revision.
How much information should be given to the contractors? There are
two main approaches to this question. On the one hand, the theorist may
want to know what would be instrumentally rational for persons given the
evidence available to them, even if that evidence does not include all the
relevant facts. For example, there are important questions regarding what
it is instrumentally rational to do given that we do not know how long we
are going to live, or whether we will be afflicted by illness or injury in the
future. Some governments require citizens to participate in compulsory
forms of collective insurance, and we may wish to know whether rational,
self-interested contractors representing all citizens would agree to such
schemes. But sometimes, as theorists, we want to know what rational and
self-interested agents would do if they had access to all the facts. For
example, we might want to know exactly what employment contracts an
employer and employee would accept if each party knew exactly what its
alternatives were. It is often valuable to know how fully informed and
rational contractors would act, even if their real-world counterparts are
neither fully informed nor fully rational.
Finally, how does rationality contractualism handle the issue of the
non-agreement point? Because the aim is often to understand what bargain
or agreement self-interested agents would reach, it’s typically crucial that
the non-agreement point be specified by reference to how well-off persons
would actually be in the absence of agreement. In this way rationality
contractualism differs sharply from fairness contractualism. In the latter,
the non-agreement point can be stipulated in any way the theorist chooses
if doing so serves the goal of making the resulting agreement fair. But
stipulating unrealistic non-agreements points will defeat the central pur-
pose of most forms of rationality contractualism, namely, discovering
what it would be rational for some group of people to accept given the
actual alternatives. Of course, we can stipulate an unrealistic or false non-

3
For a detailed account of moderate idealization, see Gaus (2011: 276–91).
82 Chapter 5

agreement point, and ask what is rational for contractors to do in light of

this unrealistic non-agreement point, but the reasons to pursue this
strategy (setting aside the aims of fairness contractualism) are unclear.

4 How Not to Use Contractualism

The preceding section focused on how a theorist might design
a contractualist scenario, depending on her more particular objectives.
In this section I turn to consider a few of the ways in which contractualism
might be misused.

4.1 Contractualism as an Account of the Fundamental Basis of

Wrongness or Injustice
If utilitarianism is true, it provides the fundamental and complete expla-
nation of wrongness and injustice. Some view contractualism as compet-
ing with utilitarianism in this sense – as offering a fundamental and
complete approach to injustice or moral wrongness. I think this misde-
scribes the point of contractualism.
First, it should be clear that neither consent contractualism nor ration-
ality contractualism can plausibly be presented as fundamental accounts
of moral wrongness or injustice. These versions of contractualism are
intended only to help us answer very speciﬁc questions about what people
would, or rationally should, accept under certain circumstances.
Second, even fairness contractualism (in any of its guises) does not
provide a fundamental explanation of what makes some act or rule
morally wrong or unjust. It is rather a useful tool to help us better under-
stand some moral or political puzzle in certain contexts, a tool that can
help us see an answer that we might previously have been unable to see,
and also a tool that enables us to organize some of our moral or political
ideals in a more systematic framework. But fairness contractualism does
not compete with utilitarianism as an account of the fundamental basis of
moral wrongness or injustice. Whenever some act, ϕ, is correctly judged
to be wrong, utilitarianism tells us that the fundamental basis of ϕ’s
wrongness is that it fails to maximize utility. Contractualists should not
respond by saying that the fundamental basis of ϕ’s wrongness is that
ϕ-ing would be prohibited by the relevant contractualist procedure.
Rather, they should follow Scanlon and agree that what makes some act
wrong or unjust will be the more speciﬁc features of the act, for example,
that it causes gratuitous suffering. Scanlon presents his moral contractu-
alism as an account of what wrongness is. It is an account of the property
that all wrong acts share, namely, that they would be disallowed by
Contractualism 83

a principle that suitably situated contractors could not reasonably reject

(Scanlon 1998: 13, 391). But this property isn’t what ultimately makes
something wrong.
I suggest that models of fairness contractualism are analogous to maps
of the physical terrain.4 We choose the map that seems to do the best job
making sense of the physical terrain given our current and independently
formed beliefs. Even if the map we have chosen turns out to be accurate in
every respect – and even if we subsequently adjust some of our pre-
existing convictions about distances or locations in light of the map’s
explanatory power – it is not the map itself that makes it true that the
mountain to the west is 500 miles away. Fairness contractualism is, in this
sense, like a map. We use our existing beliefs about fairness to construct
a map or theory that provides a more complete picture of what fairness (or
justice or rightness) requires.

4.2 Contractualism Where It Doesn’t Belong

A closely related mistake is the use of contractualism to handle moral or
political problems to which it isn’t suited. Consider, for example, the
following difficult question: when, if ever, is it morally permissible to kill
innocent people for the sake of achieving some greater good? To bring
things into sharper focus, consider the standard trolley case, where
a runaway trolley is going to kill five innocent people trapped on the
main tracks unless the trolley is diverted on to a sidetrack, where it will
instead only kill one innocent person. Is it permissible to turn the trolley,
and, if so, why? Contractualism is not a useful tool for addressing ques-
tions like this.5
You might think this sort of problem looks ripe for consent contrac-
tualism: if we could establish that the one person on the sidetrack would
consent to the turning of the trolley, this might explain why it’s permis-
sible to do something to him that is presumptively wrongful. Of course the
one person on the sidetrack will likely refuse consent if he knows doing so
entails his death, but perhaps the point is that he would consent ex ante,
from behind a veil of ignorance that prevents him from knowing whether
he will be a member of the five on the main track, or the one on the
sidetrack.
But this explanation does not accurately capture why it is permissible
(assuming it is) to turn the trolley. What makes it permissible to turn the

4
The ﬁnal four sentences in this paragraph are taken, with slight modiﬁcation, from Quong
(in press).
5
For a related discussion of contractualism as applied to this case, see Thomson (1990:
181–95).
84 Chapter 5

trolley is not that the one person would have consented ex ante from
behind a veil of ignorance, but rather the fact that turning the trolley can
avert a sufficiently great harm. The ‘consent’ of the one person is only
achieved by placing that person behind a veil of ignorance that deprives
the person from the most important piece of information relevant to the
decision, namely, that turning the trolley involves killing him for the
benefit of others. The consent obtained in this way is too far removed
from the interests and will of the actual person to carry the required
normative weight.
Perhaps this simply shows that the type of contractualism best
suited to handle these cases is not consent contractualism, but rather
fairness contractualism. The point, one might argue, is that under
suitably fair conditions (i.e. behind a veil of ignorance) all six people
would agree that the trolley should be turned, and so this establishes
that turning the trolley is the fair and thus permissible course of
action.
There are several serious problems with this view. First, in these
cases the central question is what it is permissible to do, and at
most, what fairness contractualism could tell us is what would con-
stitute a fair distribution of some harm. But the fact that distributing
harm in one way is fairer than the alternative is not sufficient to
conclude that imposing the harm in this way is permissible. From
behind a veil of ignorance, six people might also agree that
a surgeon should cut up a healthy patient and distribute his organs
to five other patients who are dying and in need of organ transplants
to survive, but hardly anyone thinks this is a permissible course of
action. Second, an act being fair is not necessary to establish its
permissibility. Suppose, for example, the runaway trolley can be
turned either to a sidetrack to the left, where it will kill your spouse,
or to a sidetrack to the right, where it will an innocent stranger.
Many will agree you are permitted to turn the trolley to the right
(you don’t have to flip a coin), but this is not a fair distribution
since the person on the right is given no chance to avoid being
killed. Finally, cases such as these are not analogous to standard
problems of distributive justice where fairness contractualism is
often deployed. In cases like the one being considered, each person
cannot share in the burdens and benefits. Rather, some people will
be forced to bear all the burdens while others reap all the benefits.
Though this may be the right thing to do in certain circumstances,
it’s not plausible to say that the burdens and benefits have been
fairly distributed.
Contractualism 85

4.3 Conﬂating Different Types of Contractualism

One of my central claims in this chapter is that there are different types of
contractualism, and that they are designed to address very different
questions. A significant problem in political philosophy, I believe, is the
failure to recognize this fact. Proponents and critics of contractualism are
both to blame for this state of affairs.
Consider, for example, the most influential version of contractual-
ism in our field. Rawls presents his aim as being to uncover the
principles of justice appropriate to regulate social cooperation for
mutual advantage among free and equal persons (Rawls 1999: 109,
456). His repeated references to mutual benefit and mutual advantage
have led a number of his critics to interpret his theory as, at least in
part, an account of what rational self-interested bargainers would
accept. The critics then point out that Rawls’s defence of his two
principles of justice as fairness, in particular the difference principle,
cannot be reconciled with the view that his principles are also
mutually advantageous. A political society regulated by the difference
principle may not, these critics plausibly argue, be advantageous for
many people who might be better off under a less egalitarian regime
(Barry 1989b: 241–54; Cohen 1995: 224–6).
I believe these critics misunderstand Rawls’s theory: his theory is best
understood as a pure version of fairness contractualism, and the state-
ments about mutual advantage do not refer to mutual advantage in the
sense familiar from rationality contractualism (Quong 2007: 77–82). But
the general point the critics make is sound: fairness contractualism cannot
be combined with rationality contractualism, at least where the aim is to
uncover what it would be rational for self-interested agents to accept.
A great deal of confusion in political philosophy is created by a failure
among those who use contractualism to be clear about the role the
contractualist device is designed to play.
A similar type of confusion can arise when we fail to distinguish
between consent contractualism and fairness contractualism. Suppose
a theorist argues that some presumptively wrong act, ϕ, may permissibly
be performed because suitably situated contractors, representing those
persons who would be presumptively wronged, would consent.
The success of this argument depends crucially on the contractual sce-
nario being designed in such a way that it approximates as closely as
possible the will of the real people in question. If the scenario doesn’t
do this – for example, if the contractors are given very different aims or
interests from those of the real people they represent, or if they are placed
behind a thick veil of ignorance – then critics may rightly argue that the
86 Chapter 5

‘consent’ obtained is too far removed from the will of the relevant real
people.6 This, recall, is one of the problems with an attempt to use
contractualism to explain why it is permissible to turn a runaway trolley
away from ﬁve people and towards one person.
I don’t wish to defend the extreme thesis that hybrid models of
contractualism are always unsuccessful. One reason to develop
a hybrid model would be if one’s wider metaethical or normative
beliefs entail an identity (or at least extensional equivalence)
between the different questions associated with different forms of
contractualism. For example, if one independently believed that
moral rules must be co-extensive with individual self-interest, then
the distinction between rationality contractualism and fairness con-
tractualism collapses. Similarly, if one had independent reasons to
believe that social morality is whatever set of rules real persons
(suitably idealized) have sufﬁcient instrumental reasons to endorse,
then the distinction between a version of fairness contractualism and
an expansive version of rationality contractualism blurs. But as
a general methodological principle, we do better to carefully distin-
guish the different varieties of contractualism.

4.4 Too Much or Too Little Content

A question theorists must confront when designing models of fairness
contractualism is this: how much normative content can be ‘front-loaded’
into the model?
On the one hand, there is a risk of assuming so much specific normative
content that the contractualist device no longer seems necessary.
We begin with all the detailed moral content we might have hoped to
get out of the contractualist procedure. Suppose the question we wish to
answer with our contractual device is this: what is the fairest principle to
regulate the distribution of income and wealth amongst citizens? What we
must not do is stipulate that the contractors are motivated to reject any
principles apart from some specific principle of distributive justice. Doing
this begs the question: we simply stipulate the answer to the question we
are allegedly investigating.
Brian Barry memorably accuses Rawls of making this mistake in his
initial attempt to explain why contractors in the original position would
favour a principle requiring saving for future generations. In reaching this
conclusion, Rawls assumes that the contractors care about the well-being
of those in the next generation. This, Barry scornfully says, is ‘like
6
Though it is sometimes the critic who bears responsibility for the mistake of conflating the
two forms of contractualism. See Sandel (1982).
Contractualism 87

a conjurer putting a rabbit in a hat, taking it out again and expecting

a round of applause’ (Barry 1989a: 505).
Although it is crucial to avoid begging the question, this doesn’t
mean that it is a mistake to build a good deal of normative content
into the design of the contractual scenario. Indeed, doing so is
essential. In order to design a contractual scenario where we can be
confident that the agreement amongst the contractors is guaranteed
to reflect what fairness requires, we must already have in hand some
independent ideas about fairness. The contractual device offers a way
of modelling and systematizing our existing convictions about fair-
ness. The hope is that we can begin with those convictions about
which we are most confident and settled, and use them to design
a contractual scenario that will help us reach answers about issues
where we are uncertain.
Most of us believe, for example, that: persons are moral equals – no
person’s interests matter more than anyone else’s; that harms imposed on
one person cannot simply be outweighed by benefits conferred on others
(the separateness of persons); and that sex, race, or social class are not
legitimate reasons to discriminate amongst persons. So we take these
convictions, among others, and represent them in the contractual sce-
nario. We can represent the moral equality of persons by ensuring that the
relative bargaining powers of the contractors is equalized. We can repre-
sent the separateness of persons by giving each contractor a veto. And we
might represent the irrelevance of sex, race, and social class by placing
parties behind a veil of ignorance. These are all ways in which a great deal
of normative content can be built in to the contractual scenario. But this
needn’t be problematic, provided that we don’t use, as one of our con-
sidered convictions to shape the design of the contractual device,
a particular answer to the question that the contractual device is meant
to answer.
Of course even if we do not assume a precise answer to the question that
the contractual device is intended to answer, there is still a sense in which
the answer on which the contractors converge should be entailed or be
inescapable given the normative design of the contractual scenario. Some
worry that this reveals the contractual device is a trick or sham. But this
objection is puzzling. As with any valid argument, the conclusion should
follow from the premises. Few people allege that valid syllogisms are
a sham because the conclusion must follow given the premises. There is
no reason to suppose this complaint has greater force when applied to
contractualism.
Perhaps the objection is not that contractualism is a trick or sham, but
rather that it’s an unnecessary device. If the agreement the contractors
88 Chapter 5

reach is inescapable given the normative assumptions we use to design the

contractual scenario, we should dispense with the contractual device and
simply present our argument as a series of premises leading to
a conclusion.
But this objection to contractualism is almost as puzzling. First,
there are often several ways an argument can be presented, and
I doubt there are decisive general reasons to favour one mode of
presentation over another regardless of the context. Second, contrac-
tualism is not simply a method by which an argument is presented. It’s
also a tool we can use to help us see an answer we couldn’t see before.
Just like a well-designed telescope allows us to see things we might not
otherwise be able to see, the process of imagining oneself behind the
veil of ignorance, or imagining what self-interested bargainers might
accept under certain constraints, can be a powerful heuristic. Finally,
contractualism can be more than a way of presenting an argument or
a heuristic; it can also function as a theoretical framework. That is,
once we have designed a contractual device that we are confident has
successfully helped us answer a particular question, we can use the
device to answer further questions, provided those questions are suffi-
ciently similar to our original enquiry. The more explanatory power
a particular form of contractualism seems to have with regard to
a normative domain, the more we can justifiably be confident that it
offers a useful framework for answering similar questions in that
domain. We may even, as Rawls suggests, sometimes choose to revise
our existing beliefs in light of the general explanatory success of the
contractualist model.

5 Summary: How to Use Contractualism

In closing, what follows is a list of some of this chapter’s main claims,
presented as a set of ‘how-to’ recommendations for using contractualism
in moral and political theory.
1) Be clear about the nature of your question. The nature of your
question will dictate which type of contractualism (consent, fairness,
rationality) you need to use.
2) Do not conflate different types of contractualism. The aims of
consent, fairness, and rationality contractualism are sufficiently dis-
tinct that they typically cannot all be successfully pursued within
a single model.
3) Having identified the type of contractualism you need, design your
model of contractualism with the following five questions in
mind:
Contractualism 89

i. Constituency: Who do the contractors represent?

ii. Primary Question: What question do the contractors confront?
iii. Motives and Interests: What are the motives and interests of the
contractors?
iv. Information: How much information should be provided to the
contractors?
v. Non-agreement: What will happen to each contractor in the event of
non-agreement?
4) Contractualism cannot provide an account of the fundamental
basis of moral wrongness or injustice. It is a model or framework
for organizing moral ideas.
5) Contractualism cannot be used to solve all moral or political
problems. It is, for example, ill-suited to helping answer questions
about so-called lesser evil justifications.
6) Models of fairness contractualism are redundant when too
much normative content is assumed – when the answer being
sought is assumed in the design of the model. But models of fairness
contractualism can also fail when too little moral content is
built in. Moral conclusions cannot be derived from a model unless
the model accurately reflects a sufficient set of prior moral convictions.

Acknowledgements
I am very grateful to Adrian Blau and Rebecca Stone for comments on
earlier versions of this chapter.

References
Barry, Brian, 1989a. Democracy, Power, and Justice. Clarendon Press.
Barry, Brian, 1989b. Theories of Justice: A Treatise on Social Justice, Vol. I. University
of California Press.
Cohen, G. A., 1995. Self-Ownership, Freedom, and Equality. Cambridge University
Press.
Dworkin, Ronald, 1973. ‘The original position’, The University of Chicago Law
Review 40: 500–33.
Gaus, Gerald, 2011. The Order of Public Reason: A Theory of Freedom and Morality in
a Diverse and Bounded World. Cambridge University Press.
Gauthier, David, 1987. Morals by Agreement. Oxford University Press.
Hobbes, Thomas, 1996. Leviathan, ed. Richard Tuck. Cambridge University
Press.
Nozick, Robert, 1974. Anarchy, State, and Utopia. Basic Books.
Quong, Jonathan, 2007. ‘Contractualism, reciprocity, and egalitarian justice’,
Politics, Philosophy, & Economics 6, 75–105.
90 Chapter 5

Quong, Jonathan (in press). ‘Consequentialism, deontology, and distributive jus-

tice’, in Serena Olsaretti, ed., The Oxford Handbook of Distributive Justice. Oxford
University Press.
Rawls, John, 1999. A Theory of Justice: Revised Edition. Oxford University Press.
Reiman, Jeffrey, 2007. ‘Being fair to future people: the non-identity problem in the
original position’, Philosophy & Public Affairs 35: 69–92.
Sandel, Michael, 1982. Liberalism and the Limits of Justice. Cambridge University
Press.
Scanlon, T. M., 1998. What We Owe to Each Other. Belknap Press.
Thomson, Judith Jarvis, 1990. The Realm of Rights. Harvard University Press.
6 Moral Sentimentalism

Michael L. Frazer

1 Introduction
Scholars across a variety of disciplines have come to believe that eighteenth-
century luminaries such as David Hume and Adam Smith were correct to
see our normative judgements as sentiments – that is, as integrating both
affective and cognitive elements, both emotions and beliefs.
There are several ways to defend this thesis (Frazer 2013: 21–4).
Psychologists, neuroscientists, and other social scientists have gathered
considerable evidence that normative judgements depend on emotions
(e.g. Damasio 1994; Marcus et al. 2000; Haidt 2001; Westen 2007).
Philosophers have then explored the deeper implications of these empiri-
cal discoveries (e.g. Nichols 2004; Prinz 2007). Normative ethicists and
political theorists have argued that emotional engagements ought to be
appreciated as positive features of our ethical and political lives (e.g.
Walzer 2004; Hall 2005; Krause 2008; Frazer 2010; Slote 2010).
During the first half of the twentieth century, many analytic metaethicists
even argued that the very concept of affectless evaluation is incoherent
(Ayer 1936: 102–19; Stevenson 1944). Under a particularly strong ver-
sion of this metaethical view – often defended under such names as
emotivism, non-cognitivism or expressivism – moral judgements consist
only of emotion, and contain no cognitive content whatsoever. Among
metaethicists today, moderate, qualified, hybrid or ‘neosentimentalist’
views are more popular (e.g. Gibbard 1990; D’Arms and Jacobson 2000).
Given the long-standing centrality of emotion in analytic metaethics, it
is surprising that sentimentalism is only now gaining a foothold in Anglo-
American political theory. Some might see this as evidence of the irrele-
vance of metaethics to politics. Even though the view that metaethics is
normatively neutral has not been popular for nearly half a century
(Gewirth 1968, 1970; Solomon 1970), the view that normative political
theory ought to be metaethically neutral is still widespread. Following
Rawls (1996: 12–15), metaethics is often classified alongside religion,
metaphysics, and even comprehensive theories of normative ethics as an

91
92 Chapter 6

inappropriate starting point for political theorizing in a democratic society

characterized by a broad diversity of competing worldviews.
Yet although there may be good reasons to try to build our political
theories on as non-controversial a philosophical foundation as possible,
our views about the necessary or proper form of normative judgements
will inevitably shape the techniques we use to try to form better ones.
The main implications of metaethics for political theory are thus likely to
be methodological rather than substantive. Most important for our pre-
sent purposes, the standard tools of analytic political theory discussed
elsewhere in this volume, which are designed to help us formulate sound
and valid arguments rather than to hone our emotional sensitivity, typi-
cally fail to speak to our sentiments.
The tension between the aspiration to make the standard methods of
analytic political theory metaethically neutral and the likelihood that their
appeal depends on some form of moral cognitivism can be felt throughout
the present volume. In Chapter 3 for example, Brownlee and
Stemplowska argue that ‘what matters in thought experiments are not
(or not exclusively) raw affective states’. Despite appearances to the con-
trary, they claim that this view is compatible with a ‘sophisticated version
of non-cognitivism’, one that ‘accepts that, even if moral judgement is
ultimately a matter of affective states, there is nonetheless a plausible
distinction to be drawn between raw affective states and “gardened” or
reflective ones’. And it is true that sophisticated sentimentalists have
rarely thought that one ought to follow one’s immediate moral feelings,
but instead must undertake a process of reflective self-correction in which
all one’s moral sentiments are progressively put under the test of their own
evaluative scrutiny (Frazer 2010).
Later in their chapter, however, Brownlee and Stemplowska defend the
permissibility of what they call ‘imaginatively opaque’ thought experi-
ments – scenarios in which it is impossible to engage empathetically with
the impossible situations in which the protagonists find themselves – on
the grounds that ‘experiments are not “run” (simply) to establish how we
(the experimenters) would feel, but to establish what we may plausibly
think’. Purely cognitive consideration of characters with whom empathy is
impossible may allow for the formation of proper normative judgements if
evaluation is a matter of pure reason, but this imaginative opacity will
prevent proper judgement under most sentimentalist theories.
If the standard methods of analytic normative theory are indeed less
attractive under a sentimentalist theory than they are under a cognitivist
one, then the committed sentimentalist is left with at least three options.
First, one might maintain that normative evaluation cannot be a legiti-
mate part of philosophy. Early-twentieth-century emotivists believed that
Moral Sentimentalism 93

only analytic methods qualiﬁed as philosophical, and that these methods

could never vindicate normative claims. Of course, considerations of pro-
fessional survival make this option rather unattractive to normative theor-
ists today. Fortunately for us, in the second half of the twentieth century
normative moral and political philosophy came to establish themselves as
important subfields in even the most analytic philosophy departments.
Second, the standard techniques of analytic normative theory could be
reinterpreted in ways that make them a better fit with sentimentalism.
Even if most analytic philosophers believe that they are only in the busi-
ness of constructing sound and valid arguments, they may nonetheless
succeed in improving their readers’ normative judgements by encoura-
ging psychologically holistic self-scrutiny, sparking empathy with other’s
emotional experiences, and increasing sensitivity to the affectively salient
features of collective life. I have argued elsewhere that Rawls (1999)
succeeds in engaging our moral sentiments in exactly these ways, and
that his theory of justice is far more sentimentalist than Rawls himself ever
acknowledges (Frazer 2007). Many of the most successful pieces of moral
and political philosophy operate similarly; just think of the emotional
power of Singer (1972) equating a nearby child drowning in a mud
puddle with one dying from famine in Bangladesh. Yet since most affec-
tive effects of analytic philosophy are unintended by-products of its usual
methods, this alternative is unsuited for a chapter designed to provide
a guide to the intentional practice of sentimentalist theory.
Only sentimentalists who are willing to violate the conventions of
contemporary philosophy will deliberately seek to evoke moral senti-
ments in their readers. As such, my focus here will be on a third alternative
available to sentimentalists, who can make common cause with all those
seeking to bring a greater diversity of methods and approaches to the
practice of Anglo-American political theory.
Once we accept that political theory can be more than a matter of applied
logic, we can consciously develop techniques designed to spark psychologi-
cally holistic reflection in our readers. Section 2 of this chapter will discuss
the implications that sentimentalism should and should not have for our
approach to normative political theory. Section 3 defends the permissibility
of this theoretical approach against possible objections. Section 4 concludes
by drawing on the work of Hume and Smith to provide concrete advice
about how to apply sentimentalist techniques successfully.

2 Sentimentalist Theory and Impassioned Practice

This is hardly the place to provide a full defence of sentimentalism on
empirical, normative or conceptual grounds. Instead, let us assume, for
94 Chapter 6

the sake of argument, that sentimentalism in some form is true, or at least

more plausible than the available alternatives. The obvious methodolo-
gical response would seem to be an impassioned form of normative
argumentation. Nussbaum, for one, suggests that conducting
sentimentalist theory dispassionately gives the appearance of a ‘peculiar
sort of self-contradiction between form and thesis’. Consider, she sug-
gests, an article that
argues that the emotions are essential and central in our efforts to gain under-
standing on any important ethical matter; and yet it is written in a style that
expresses only intellectual activity and strongly suggests that only this activity
matters for the reader in his or her attempts to understand. There might have been
some interesting reason for writing this way; but usually, in cases of this kind, the
whole issue had just not arisen. Such articles were written as they were because
that was the way philosophy was being written, and sometimes because an
emotive or literary style would have evoked criticism, or even ridicule.
(Nussbaum 1990: 21)

Nussbaum’s argument would seem to suggest that there is at least a pro

tanto reason for those convinced of the truth of moral sentimentalism to
adopt different approaches to philosophical argumentation than those
their cognitivist colleagues espouse, one stemming from the desire to
avoid contradiction between form and content. In order to avoid such
contradictions, it can be argued that diverse philosophical positions
require different investigative methods and modes of expression
(Stewart 2013: 1–12, 159–70).
In fact, however, there is neither any logical entailment nor a strong
empirical correlation between a commitment to moral sentimentalism
and the adoption of a distinctively impassioned approach. Sentimentalists
often draw a distinction between the proper practice of philosophy and
the proper practice of everyday reflection, maintaining that it is not
appropriate for philosophers to express or evoke moral sentiments, but
only to discover the truth about them. As a result, one might argue that
while first-order practical reflection is and ought to be
impassioned, second-order philosophical investigation into the nature
of this first-order reflection should not be.
There is considerable precedent for this sort of unemotional analysis of
human emotion. While ancient philosophers may have disagreed about
the proper place of passion in moral life, most agreed that the study of
emotion was an important part of the philosophical enterprise, such as in
the fields of rhetoric and poetics. Yet Aristotle’s Rhetoric and Poetics are as
dry as anything else in his corpus. Aristotle carefully examines the evoca-
tion of human emotion while failing to evoke any significant feelings in the
reader (with the possible exception, depending on one’s predilections, of
Moral Sentimentalism 95

intellectual excitement or boredom). The same dispassion can be found

in the study of emotion by most experimental psychologists and senti-
mentalist metaethicists today. Hume is therefore guilty of no inconsis-
tency when he insists in the Treatise that, when it comes to practical
decision-making, ‘reason is and ought only to be the slave of the passions’
(2000a: 2.3.3.4, 226), while also complaining that too often in philosophy
‘it is not reason, which carries the prize, but eloquence’ (2000a:
Intro.2, 3).
At the conclusion of the Treatise, however, Hume makes clear that
‘were it proper in such a subject to bribe the reader’s assent, or employ
anything but solid argument’, his seemingly dry anatomy of the moral
sentiments is in fact ‘abundantly supplied with topics to engage the
affections’ (2000a: 3.3.6.3, 394). In his later writings, Hume lost this
reticence about engaging readers’ affects, and used a wide variety of
literary techniques in the attempt to do so (Frazer 2015).
Hume came to see that, if his readers fail to feel affective approbation
for the moral sentiments that he is analysing, his normative defence of
these sentiments becomes merely descriptive. Since all evaluation con-
tains an affective component, sentimentalist theory cannot be consis-
tently normative without being impassioned. While there is no
necessary connection between sentimentalism as such and the practice
of impassioned theorizing, the conjunction of sentimentalism with
a commitment to practising normative evaluation does require emotion.
If sentimentalists abide by the standard methods of analytic philoso-
phy, they will only evoke affective approbation and disapprobation unin-
tentionally, if at all. Their work may end up merely descriptive. Even
worse, it might actually weaken our moral sentiments. Hume is deeply
concerned that any moral sentiment can ‘with facility, be reﬁned away . . .
in sifting and scrutinizing it, by every captious rule of logic, in every light
or position, in which it may be placed’ (Hume 1985: 482). To be sure,
Hume holds that those who tamp down their moral sentiments in this way
are indulging in a ‘false philosophy’, but even accurate philosophy may
have similar results.
Although the minute moral distinctions that casuists draw are often
unobjectionable when considered individually, the very process of
coldly drawing distinction after distinction seems to desensitize both
casuists and their audience to the properly affective features of moral
life. Anyone who has ever sat through a dispassionate – or even down-
right cheerful – discussion by analytic ethicists of scenarios involving
running over people with trolleys and resorting to cannibalism on life-
boats has to worry that standard analytic methods can have precisely this
effect. Nor is this problem a new one: Seneca also objected to the logic-
96 Chapter 6

chopping playfulness of the ethicists of his own day. ‘It makes one
ashamed’, he writes, ‘that men of our advanced years should turn
a thing as serious as this into a game’ (Seneca 2004: 97).

3 Objections to Impassioned Philosophy

Before we can discuss practical techniques for successfully evoking moral
sentiments in our readers, we have to address the argument that this is not
an appropriate goal to pursue in the ﬁrst place. I will focus on two
important objections to the practice of impassioned philosophy: the
objection from disciplinary distinctions, and the objection against
manipulation.

3.1 The Objection from Disciplinary Distinctions

There is a remarkable degree of methodological conservatism in most
academic disciplines. The most common criticism of anyone violating
local conventions is that their work is not ‘real’: not ‘real philosophy’ or
‘real political theory’. If research resembles work done outside the acad-
emy rather than in an adjacent discipline, it may not even qualify as ‘real
scholarship’.
While the professional norms of each discipline may seem self-
justifying to most of their practitioners, Applebaum (2000) observes
that what creates an ethical justification for abiding by the duties of
a particular profession is the importance of that profession for one’s larger
society. If a particular vocation serves no justifiable purpose, then the
internal rules of that profession are not a legitimate branch of any larger
ethical system. We’re all better off without professional gladiators and
professional torturers. The question is therefore whether the reigning
standards of a given profession – such as the methodological norms of
an academic discipline – aid in the achievement of a genuinely valuable
social purpose.
The methodological standards governing analytic philosophy are
less than a century old. The idea of philosophy as a distinct discipline
is slightly older. In the eighteenth century and earlier, philosophy was
simply the search for general, nomothetic truth. As such, it included
much of what we now call ‘science’, but still excluded activities that
have something other than general truth as their aim – such as
persuasive oratory and idiographic narrative, whether factual or
fictional.
The principle that the search for general truth precludes the evocation
of emotion has long had its defenders. Recall the young Hume’s view that
Moral Sentimentalism 97

when discussing ‘such a subject’, it is inappropriate to ‘employ anything

but solid argument’.
Later, Smith contrasts three types of writing: narrative, didactic and
rhetorical. Didactic discourse, of which philosophical writing is the para-
digm, ‘proposes to put before us the arguments on both sides of the
question in their true light, giving each its proper degree of influence,
and has it in view to persuade no farther than the arguments themselves
appear convincing’. By contrast, rhetorical discourse ‘endeavors by all
means to persuade us, and for this purpose it magnifies the arguments on
the one side and diminishes or conceals that might be brought’ on the
other. The goal of a didactic discourse is primarily instruction, and only
secondarily persuasion; the goal of rhetoric is persuasion and ‘instruction
is considered only so far as it is subservient to persuasion, and no farther’
(Smith 1985: 12, i.149–150, 62).
Although the line between rhetorical and didactic compositions is
sharply drawn, the line between didactic and narrative writing is blurrier.
This is most evident in the genre of history, a form of narrative writing that
incorporates significant didactic elements. The didactic power of histor-
ical and other factual narratives, moreover, suggests a sort of hybrid genre
in which an author establishes ‘certain principles’ that are ‘confirmed by
examples’ (Smith 1985: 17 ii.17, 90–1). This hybrid approach is particu-
larly well-suited for what Smith calls ‘the practical sciences of politics and
morality or ethics’, which he complains ‘have of late been treated too
much in a speculative manner’ (Smith 1985: i.102, 41).
If ethics and politics are treated too speculatively – that is, only in terms
of nomothetic principles rather than in terms of concrete events in the
lives of particular human beings – our moral sentiments may be blocked.
‘When a philosopher contemplates characters and manners in his closet’,
Hume protests, ‘the general abstract view of the objects leaves the mind so
cold and unmoved, that the sentiments of nature have no room to play,
and he scarce feels the difference between vice and virtue’. A historian, by
contrast, ‘places the objects in their true point of view’, and hence devel-
ops ‘a lively sentiment of blame and praise’ (Hume 1985: 568). In all of
his post-Treatise writings, Hume is therefore careful to illustrate every
moral and political point with concrete examples. Hume usually takes his
examples from history and the classics; Smith tends to prefer everyday
experiences that his readers will recognize from their own lives.
Interdisciplinarity has thus been a hallmark of moral sentimentalism
from its beginning (Frazer Forthcoming). Moral sentiments are unlikely
to be evoked effectively when one strictly follows the norms of didactic,
speculative eighteenth-century philosophy, let alone the norms of analytic
philosophy in the twenty-first century. As such, sentimentalists can make
98 Chapter 6

common cause with those practising what is now called realist or non-
ideal theory, overcoming the barriers between political philosophy and
social science (see Jubb, Chapter 7 in this volume).
Much of the empirical work realists cite today is the sort of nomo-
thetic social science that still qualifies as philosophy in the eighteenth-
century sense, the kind of ‘general facts about human society’ that
Rawls allows to the otherwise ignorant agents in the original position
(Rawls 1999: 116). In order to integrate narrative and didactic ele-
ments in the manner Hume and Smith advocated, sentimentalists are
more likely to draw on idiographic work in history and ethnography.
They are particularly likely to be drawn to the work of social scientists
who favour interpretive rather than causal explanations of human
behaviour. In the interpretive tradition, tracing others’ narratives
allows us to identify their motivations, values and worldviews, hence
achieving empathetic understanding, what Weber (1978: 7–8) calls
Verstehen. Although Weber believed that Verstehen is compatible with
a commitment to value-neutrality, sentimentalists seek such empa-
thetic understanding insofar as it allows for accurate, affectively
laden normative judgements.
There is a growing awareness in the social sciences that accurate
accounts of particular incidents can serve as what Thacher (2006) calls
‘normative case studies’. A ‘phronetic’ social science of the sort Flyvbjerg
(2001) advocated considers such cases in order to help us develop the
skills of ethical judgement required to guide public decision-making for
the better. It is a friendly amendment to this agenda to understand the
ethical skills being developed to consist largely in emotional, empathetic
sensitivity.
In addition to joining forces with those seeking to challenge the
division between normative theory and social science, sentimentalists
may also ally themselves with those opposed to the division between
philosophy and the rest of the humanities – especially the movement
that Danto (1985: 63) calls ‘philosophy as/and/of literature’. To be
sure, turning from factual to fictional narratives poses the danger that
fictions will be designed manipulatively, artificially evoking inap-
propriate moral sentiments. While the particular challenges that
a sentimentalist faces when drawing on imaginative literature will be
discussed later in this chapter, this worry suggests a second objection
to the sentimentalist approach. Even if it is acceptable to challenge
reigning disciplinary distinctions, one might still worry that it is unac-
ceptable to evoke moral sentiments in one’s readers because doing so
is manipulative.
Moral Sentimentalism 99

3.2 The Objection Against Manipulation

Recall Hume’s contention that evoking moral sentiments is an attempt to
‘bribe the reader’s assent’ (Section 2). On this view, evoking a reader’s
sentiments might itself be ethically objectionable – a deformation of the
appropriate relationship between an author and a reader, just as bribery is
a deformation of the proper relationship between a citizen and an official.
This moral objection holds regardless of an author’s disciplinary identity,
or lack thereof.
Bentham was later to sharpen the point still further, arguing that what
Hume said was merely bribery is in fact ‘a cloak, and pretense, and
alignment, to despotism’. Emotionally evocative literary effects are but
‘so many contrivances for avoiding the obligation of appealing to any
external standard, and for prevailing upon the reader to accept of the
author’s sentiment or opinion as a reason for itself’ (Bentham 1962: 2:14,
8–9).
Bentham foreshadows the hostility to evoking emotion that we find in
more recent liberal theories. Darwall (1995: 74) interprets Bentham as
regarding an obligation to provide public justification ‘as a necessary
condition of liberal public moral debate. When people make a moral
claim on others, he suggests, but are unwilling to offer a reason for
doing so that others could be expected to accept without already sharing
those moral views, this is implicitly coercive.’ The contrast here is
between the power of an authoritarian demagogue – for whom emotion-
ally manipulative propaganda is just another means of control, no differ-
ent in this respect from the secret police – and the non-coercive exchange
of reasons among equals in democratic discourse.
There is, however, a regrettable slippage in this argument between
the idea that public justification requires the provision of public reasons
(in the sense of grounds for justification which you expect the entire
community to share) and that it requires reasoning (in the sense of
collective reliance on the rational faculty alone). As recent sentimentalist
political theorists have pointed out, an important implication of moral
sentimentalism is that our reasons – whether individual and private or
public and shared – cannot come from reason alone (Krause 2008; Frazer
2010; Kingston 2011). Shared reasons imply the existence of shared
feelings; this is why sympathy or empathy, the faculties by which emotions
are communicated from one person to another, are so important in most
versions of sentimentalism (but see Prinz 2011). It is no more despotic,
coercive or manipulative for members of a political community to share
emotions with one another than it is for them to provide rational argu-
ments to one another.
100 Chapter 6

Recent philosophers have struggled to provide a precise account of

what constitutes manipulation, but there is a general consensus that
evoking emotion is neither necessary nor sufficient for manipulation to
occur. Most see manipulation as influence that is at best indifferent to
whether its target adheres to the relevant standards of good judgement
(Coons and Weber 2014: 11–14). This raises the question of whether
these relevant standards are objectively binding or are grounded in the
subjective commitments of the manipulator, the target, or both.
Regardless, if all the parties involved are committed sentimentalists –
and if they are objectively correct in being so – then the ideals in question
will call for feeling proper emotions. This is not to say that morally
unacceptable forms of interpersonal manipulation are impossible
among sentimentalists, but that they are just as likely to involve sophistic
argumentation as they are to involve the artificial evocation of affect.

4 Techniques for Impassioned Philosophy

Simply establishing that it is permissible to evoke moral sentiments in our
readers would be of little value were it impossible for us ever do so
successfully. Fortunately, this sentimentalist goal is eminently achievable.
Admittedly, evoking emotion is an art rather than a science, and using the
techniques described below is neither necessary nor sufﬁcient for impas-
sioned philosophy. It is not necessary because, as has already been men-
tioned, analytic philosophers often end up evoking moral sentiments
without intending to do so. It is not sufﬁcient because, if not practised
with the right degree of philosophical and literary artistry, these techni-
ques cannot be guaranteed to succeed.

4.1 Write Interesting Stories

As is now already evident, the best format for impassioned, sentimentalist
normative theory is a genre combining both idiographic and nomothetic
elements, both particular narratives and general philosophy. The most
important thing sentimentalists should keep in mind when writing a story
is that it should be interesting, both in the sense that we use the term
‘interesting’ today and in its original meaning, still dominant in the eight-
eenth century, of involving or engaging our interests.
The antiquated and the current uses of ‘interesting’ are closely con-
nected. As Vermeule (2010: 41) observes, engaging our empathy with
another human mind is a sure-ﬁre way to hold our attention.
Sentimentalist theory should be ﬁlled with interesting stories,
ones in which we sympathetically engage with the characters,
Moral Sentimentalism 101

allowing their needs and interests to become our own. Without this
engagement with another’s interests, we are likely to ﬁnd ourselves left
both cold and bored.
Livy, for example, interests us because ‘we enter into all the concerns of
the parties and are almost as much affected with them as if we ourselves
had been concerned in them’ (Smith 1985: 17, ii.27, 95–6). Via sym-
pathy, their interests become our own. Livy and Tacitus lead us ‘so far
into the sentiments and mind of the actors that they are some of the most
striking and interesting passages to be met with in any history’ (Smith
1985: 20, ii.67, 113).
Well-written history is always interesting in this way, as are philosophi-
cal, didactic writings that make frequent use of well-written narratives.
This is probably one of the reasons historians and ethnographers still
regularly ﬁnd an audience among the educated reading public, while
philosophers who are not also storytellers rarely do so.

4.2 Show, Don’t Tell

Manipulative rhetoricians often succeed in arousing moral sentiments in
their audiences, but they do so in ways very different from those writing in
the mode that Hume and Smith advocate. The author of a well-written
historical narrative, Smith points out, ‘may excite grief or compassion but
only by narrating facts which excite those feelings; whereas the orator
heightens every incident and pretends at least to be deeply affected by
them himself’ (Smith 1985: 18, ii.38, 101). A good historian ‘acts as if he
were an impartial narrator of the facts’ (Smith 1985: 7, i.82–3, 35), and
‘exclamations in his own person would not suit with the impartiality he is
to maintain and the design he is to have in view of narrating facts as they
are without magnifying or diminishing them’ (ibid., 18, ii.40, 101).
The best writers in other genres, including the hybrid genre that is most
appropriate for sentimentalist political theory, typically share the histor-
ian’s characteristic ‘modesty’ in this regard. When detailing their nar-
ratives, sentimentalists should refrain from telling their readers
how to feel, letting their stories speak for themselves.
Smith and Hume were both, like many of their time, admirers of the
essayist Joseph Addison. While previous print moralists would harangue
their readers with moral exhortation, following the model of the church
sermon, Addison prefers to ‘deliver his sentiments in the least assuming
manner; and this would incline him rather to narrate what he had seen
and heard than to deliver his opinions in his own person’ (Smith 1985: 10,
i.128, 53). Addison’s authorial persona, like Smith’s ideal moral judge, is
an impartial spectator – ‘Mr. Spectator’, as Addison calls him – who
102 Chapter 6

simply tells it like he sees it. The immense popularity and inﬂuence of
Addison’s Spectator is evidence of the power of such straightforward
writing.
It is a striking fact of human psychology that this modest approach can
arouse even greater emotional reactions than those aroused by direct
appeals. Smith argues that it is a ‘general rule that when we mean to
affect the reader deeply we must have recourse to the indirect method of
description, relating the effects the transaction produced both on the
actors and the spectators’ (Smith 1985: 16, ii.7, 86–7). Creative writers
are taught the rule ‘show, don’t tell’, not to establish impartiality or to
avoid manipulation, but simply to maximize emotional impact. The fact
that the same technique is the best way to achieve these two very different
objectives is a happy coincidence of human psychology.

4.3 Focus on Particulars, then Generalize

While analytic theorists often try to deduce particular conclusions from
general moral principles, Smith insists that the general rules of morality
should only be derived inductively. Specifically, they are to be ‘formed, by
finding from experience, that all actions of a certain kind, or circum-
stanced in a certain manner, are approved or disapproved of’ (Smith
1984: III.4.8, 159). Sentimentalists should begin with empathetic
consideration of particulars, only then attempting to extrapolate
from these cases to more general principles, rather than vice
versa. When a general principle is stated in the abstract, we are unlikely
to have any emotional reaction. When, however, it is inductively derived
from a series of cases, each one of which evoked our moral sentiments, it
can carry with it the force of the cases that serve as its foundation.
Even within a particular normative case study, the reader’s attention is
best focused on the subjective experiences of particular individuals. Smith
says Tacitus is a particularly good example of an author who focuses on
‘internal’ micro-level phenomena rather than ‘external’ macro-level ones.
When Tacitus fixes our attention on the psyche of a single individual, our
sympathies are ‘as it were concentrated, and become greatly stronger than
when separated and distracted by the affecting circumstances that befell
the several persons involved in a common calamity’ (Smith 1985: 20,
ii.66, 113).
Once we understand the individual experience of a certain event, we
can then generalize to the effects of that event on relevantly similar others.
It is then the empathetic understanding of particular experiences that
allows us to inductively derive general principles about similar cases from
the full range of human life.
Moral Sentimentalism 103

It is also the empathetic understanding that each person in a statistic

represents an entire subjective world of experience that can prevent large
numbers from having their usual morally distorting effects. Bertrand
Russell is alleged to have said that ‘the mark of a civilized man is
the capacity to read a column of numbers and weep’, but as Vermeule
(2010: 33) observes, ‘on this view, none of us is really civilized’. After we
have empathetically experienced the suffering of a single individual,
however, learning that a troublingly large number of others have experi-
enced something similar can have a profound emotional effect. When we
are discussing global poverty, for example, we must never forget that we
are discussing the lives of millions of real human beings, each of which
could be recounted with the novelistic detail of the stories in Boo’s (2012)
account of a single Mumbai slum.
Guenther’s (2013) phenomenological study of solitary conﬁnement is
a good model for how to combine Boo’s focus on individual experiences
with the pursuit of general philosophical conclusions. Guenther helps us
understand that the experience of years of forced isolation is the phenom-
enological equivalent of death. Cut off from all others, a prisoner ceases to
have the inescapably social experiences that are the markers of living
humanity. When we empathetically feel the horror of such living death,
we understand the cruelty of subjecting even a single individual to it.
We can then extrapolate to the full moral signiﬁcance of subjecting
thousands upon thousands of prisoners to such an experience throughout
a nation’s penal system. We are then also able to begin establishing
normative principles about the importance of social interaction to
human life more generally.

4.4 Consider Multiple Perspectives

Whenever we seek to generalize from the subjective experience of a single
individual, there is a danger that our generalization will fail to capture the
experiences of others. If our chosen individual is not typical, then our
generalization may prove misleading. Although we lack the imaginative
and emotional resources to engage empathetically with a statistically
representative sample of any large population, we can make a concerted
effort to ﬁnd counterexamples to any general theses, and then give the
individuals involved in these counter-cases our full empathetic considera-
tion. If the most likely potential counterexamples turn out to have experi-
ences relevantly similar to those we have already considered, this provides
important evidence in favour of our general hypothesis. Those wishing to
question our generalization are always free to suggest other
counterexamples.
104 Chapter 6

In many politically important cases, however, generalizations that hold

true with regard to the subjective experiences of all individuals involved are
neither possible nor desirable. Politics typically involves conflict between
competing interests and worldviews. When groups with radically dif-
ferent experiences of a given situation are in conflict with each
other, it is appropriate to consider the perspectives of representa-
tive members of each of the conflicting groups. The ideal moral
judge, Smith famously argues, is an impartial spectator; in cases where
we are naturally biased in favour of one party in a conflict, our moral
sentiments will be improved if we try to overcome our initial partiality.
Smith argues that, when observing a conflict, an impartial spectator
must attempt to achieve what he calls ‘divided sympathy’ (Smith 1984: I.
ii.3.1, 34). We have to consider the experiences of those on both sides of
the conflict empathetically, and try to form sensitive evaluations of each of
them in turn. Doing so cannot involve a single narrative with a single
protagonist, but multiple, Rashomon-style narratives of a given case in
which the perspective of each party is given due consideration.
Academics tend to come from the political left (Gross 2013), and
typically have an inclination to side with those whom they see as
oppressed. But Smith warns us against this instinct, and urges us to
consider the perspective of both the alleged victim and the alleged oppres-
sor in any potential case of injustice. What our initial reaction tells us is
cruel or unjust may, with greater empathetic consideration of all the
parties involved, turn out to have been motivated by laudable moral
sentiments, such as a commitment to the common good (Smith 1984:
II.ii.3.8, 88–9). Neither the perspective of the powerful nor that of the
powerless should be privileged. We may ultimately conclude that some
current practice – modern policing, for example – is indeed oppressive,
but doing so impartially requires that we consider the point of view of the
police (as in Fassin 2013) as well as the policed (as in Goffman 2014).

4.5 Use Psychologically Realistic Fiction

Sentimentalist normative argumentation requires piling narrative upon
narrative, but narratives are an expensive resource. All consideration of
narratives is emotionally and imaginatively draining – especially when
contrasted with the ease with which a clever undergraduate can construct
sound and valid arguments. Any adequate narrative will demand more
space than is typically possible in a journal article. And some kinds of
narrative are also expensive to source, requiring years of ethnographic
ﬁeldwork or archival exploration. This raises the question of whether the
narratives that sentimentalists use must be drawn from the careful, fact-
Moral Sentimentalism 105

checked work of historians, ethnographers and journalists, or can they

just be made up – whether by theorists themselves or by professional
writers of fiction?
Hume himself took a clear stand against creative writing. While poets
‘can paint virtue in the most charming colours’, their undisciplined
imaginations often lead them to ‘become advocates for vice’.
By contrast, ‘historians have been, almost without exception, the true
friends of virtue, and have always represented it in its proper colours’
(Hume 1985: 567). While it would be wrong to conclude that historians,
ethnographers and journalists have never been guilty of emotionally
manipulating their readers through the use of one-sided, unrealistic or
even wholly fabricated narratives, when they have done so they have been
violating the norms of their respective professions. Creative writers are
not held accountable to reality in this way.
Those defending fiction under the rubric of ‘philosophy as/and/of
literature’ find their inspiration in Aristotle, who maintains that ‘poetry
is more philosophical and more elevated [that is, of greater ethical import]
than history, since poetry relates more of the universal, while history
relates particulars.’ While history recounts ‘actual events’, poetry
recounts ‘the kinds of things that might occur . . . the kinds of things
which it suits a certain kind of person to say or do, in terms of probability
or necessity’ (Aristotle 1995: 1451b, 59–61). Sentimentalists are likely to
value fiction most highly in terms of its ability to trace what Nussbaum
(1995: 5) calls ‘the effect of circumstances on the emotions and the inner
world’ and its unparalleled efficacy in promoting ‘identification and
sympathy in the reader’.
Just as fiction can describe ‘the kinds of things that might occur’,
however, it can also describe the kinds of things that could never occur.
Fiction need not be realistic in all its details to be useful to senti-
mentalists, but if it is to show us anything real about the moral
sentiments, then characters’ inner worlds must be realistic, even
if the outer worlds that surround them are not. It is fine to fill
a fictional case with dragons, spaceships, experience machines and bar-
oque trolley systems, as long as we can empathize with the three-
dimensional characters trying to navigate these bizarre circumstances.
In order not to distort our moral sentiments, however, writers must
avoid flat characters like demonic villains and angelic heroes.
The presence of such impossible people is a warning sign that readers
are being manipulated.
The problem here is that, while it is always possible to check the
accuracy of a factual narrative, there is no objective test for the psycholo-
gical realism of a fictional narrative. All we can rely on is our own power to
106 Chapter 6

recognize and empathize with what we appreciate as plausibly human

states of mind. For example, it seems psychologically realistic to most
readers that the protagonist of Thomson’s (1971) famous thought experi-
ment about abortion would deeply resent being attached to an ailing
violinist for nine months. Since no one has actually experienced this
particular procedure, however, there is no way to check how it actually
makes someone feel.
With their preference for high culture over low, academics tend to
assume that great literature is more likely to be psychologically realistic
than popular fiction or their own amateur storytelling. Philosophers have
a particular liking for recent, philosophically-informed high fiction, such
as the works of Milan Kundera, J. M. Coetzee, and David Foster Wallace.
Nineteenth-century novels are also popular, from those by Jane Austen at
the beginning of the century to those by Henry James at the end (the latter
no doubt a favourite, at least in part, thanks to the reflected philosophical
prestige from brother William).
All of these are wonderful sources for narratives, to be sure, but literary
genius is by no means necessary for the purpose of sentimentalist norma-
tive argumentation. Smith’s own homely stories throughout The Theory of
Moral Sentiments – think, for example, of the unfortunate tale of the poor
man’s son (Smith 1984: IV.i.8, 181) – are all highly realistic examples of
the kinds of effects that circumstances can have on our inner life, but none
rise to the level of great literature. Thomson’s is a more recent example of
a successful philosopher-written narrative, as is Williams’s (1981: 18)
more realistic story of the spouse-rescuer with ‘one thought too many’.
While it is important not to underestimate the literary gifts of Smith,
Williams, or Thomson, theorists without the creative genius of Kundera
or James might nonetheless hope to craft serviceable fictions of this
calibre.

4.6 Write in Simple, Ordinary Language

Good writers, in any genre, share sentiments with their readers. The
central thesis of Smith’s Lectures on Rhetoric and Belles Lettres is that
‘when the sentiment of the speaker is expressed in a neat, clear, plain
and clever manner, and the passion or affection he is possessed of and
intends, by sympathy, to communicate to his hearer is plainly and cleverly
hit off, then and then only the expression has all the force and beauty that
language can give it’ (Smith 1985: 6, i.v.56, 25). The Lectures is by and
large a how-to guide designed to help Smith’s students achieve this goal.
Sentimentalists today would still be well advised to follow his stylistic
suggestions.
Moral Sentimentalism 107

A recurring theme in Smith’s lectures is the contrast between the

straightforward lucidity of Jonathan Swift and the opaque floridity of
the third Earl of Shaftesbury. Shaftesbury, Smith explains, fell prey to
a mistake to which we are all vulnerable. ‘The idea we form of a good style
is almost contrary to that which we commonly hear,’ Smith explains.
‘Hence it is that we conceive the further one’s style is removed from the
common manner . . . it is so much the nearer to the purity and perfection
we have in view’ (Smith 1985: 8, i.103, 42). This belief is, of course,
mistaken. In order to evoke sentiments in their readers effectively,
writers should use a style of prose continuous with the ordinary
language that is the normal vehicle of emotional communication
in everyday life.
Academics today would never be attracted to the aristocratic excesses
of Shaftesbury’s antiquated mode of writing. Instead of aping the baroque
floridity of aristocratic oratory, most philosophers now seek to gain pres-
tige by aping the dry, technical style that predominates in high-status
STEM fields. Yet there are a number of ways in which these seemingly
opposed modes of artificial communication are surprisingly similar. For
example, the tempo of normal speech varies considerably with the matter
being discussed, but both artificial modes maintain a constant rhythm
even when ‘this uniform and regular cadence is not at all proper’ (Smith
1985: 5, i.50, 22). Both artificial modes make everyone sound the same,
whereas in natural communication ‘when all other circumstances are
alike the character of the author must make the style different’ (Smith
1985: 8, i.97, 40). The list could be continued; in each case, artificial
styles prevent emotional reactions in readers while everyday language
encourages them.
Perhaps the most common way of distinguishing artificial modes of
communication from ordinary speech is through the use of foreign terms.
Greek or Latin words and phrases – and, even worse, Latin grammatical
structures – lend an unearned air of authority to English prose, but at the
price of turning it into something that is not quite English. Bad writing,
both in Smith’s day and our own, often has ‘a great deal of the air of
translations from another language’ (Smith 1985: 2, i.10, 7).
Smith argues that the words we use ‘should be natives . . . of the
language we speak in’. As always, this rule is justified in terms of the
best means of evoking sentiments in our readers. ‘Foreigners’, Smith
explains, ‘though they may signify the same thing, never convey the idea
with such strength as those [words that] we are acquainted with’ (Smith
1985: 2, i.1, 3). Just think of the effects of choosing a word like ‘ressenti-
ment’ over ‘resentment’. The gain in prestige is hardly worth the loss of
evocative force. To be sure, Smith admits that foreign words ‘may be
108 Chapter 6

naturalized by time and be as familiar to us as those which are

originally our own, and may then be used with great freedom’ (Smith
1985: 2, i.1, 3). Only then can these words connect with everyday life and
the speech found within it, and hence carry the emotional force of these
associations with lived experience.
If a neologism is needed, it should be constructed from one or
more terms found in everyday English. Smith praises Greek philoso-
phers for using only words found in their own language even when
coining new terms of art (Smith 1985: 2, i.4, 4). In later centuries,
German authors made use of a similar practice. Contrast Freud’s das
Es, das Ich and das Über-Ich, to his English translator’s id, ego, and
superego.
The effect of all of these recommendations will be to narrow the gulf
between political theory and all other forms of discourse, both elite and
popular. This loss of disciplinary distinctiveness, however, is to be cele-
brated rather than mourned. For one thing, it is likely to increase the
impact of political theory outside the academy. Lay readers are happy to
learn from academics, but only if these academics are willing to speak in
the interesting, unavoidably impassioned language of normal human life.
To be sure, public intellectuals are often excoriated for failing to produce
‘real scholarship’, but we have already seen that this accusation should
carry no ethical weight.
‘Be a philosopher’, Hume famously urges us, ‘but amidst all your
philosophy be still a man’ (Hume 2000b: 1.6, 7). Since academics
must, unavoidably, remain human beings, we should not be ashamed
that the same sort of impassioned argumentation that is persuasive among
the general population is also the best form of argumentation available
to us.

References
Applebaum, Arthur, 2000. Ethics for Adversaries: The Morality of Roles in Public and
Professional Life. Princeton, NJ: Princeton University Press.
Aristotle, 1995. Poetics, ed. and tr. Stephen Halliwell. Cambridge, MA: Harvard
University Press.
Ayer, A. J., 1936. Language, Truth and Logic. New York: Dover Publications.
Bentham, Jeremy, 1962. ‘Introduction to the principles of morals and legisla-
tion’, in J. Bowring., ed., The Works of Jeremy Bentham. New York: Russell &
Russell.
Boo, Katherine, 2012. Behind the Beautiful Forevers: Life, Death and Hope in
a Mumbai Undercity. New York: Random House.
Box, M. A., 1990. The Suasive Art of David Hume. Princeton, NJ: Princeton
University Press.
Moral Sentimentalism 109

Coons, Christian and Michael Weber, 2014. ‘Introduction: investigating the core
concept and its moral status’, in Christian Coons and Michael Weber, eds.,
Manipulation: Theory and Practice. Oxford University Press, 1–16.
Damasio, Antonio, 1994. Descartes’ Error: Emotion, Reason and the Human Brain.
New York: G. P. Putnam’s Sons.
Danto, Arthur, 1985. ‘Philosophy as/and/of literature’, in John Rajchman and
Cornel West, eds., Post-analytic Philosophy. New York: Columbia University
Press.
D’Arms, Justin and Daniel Jacobson, 2000. ‘Sentiment and value’, Ethics 110,
722–48.
Darwall, Stephen, 1995. ‘Hume and the invention of utilitarianism’, in
M A. Stewart and John Wright, eds., Hume and Hume’s Connexions. State
College: Pennsylvania State University Press, 58–82.
Fassin, Didier, 2013. Enforcing Order: An Ethnography of Urban Policing. Malden,
MA: Polity Press.
Flyvbjerg, Bent, 2001. Making Social Science Matter. Cambridge University Press.
Frazer, Michael, 2007. ‘John Rawls: between two enlightenments’, Political Theory
35, 756–80.
Frazer, Michael, 2010. The Enlightenment of Sympathy: Justice and the Moral
Sentiments in the Eighteenth Century and Today. Oxford University Press.
Frazer, Michael, 2013. ‘Sentimentalism without relativism’ in James Fleming, ed.,
NOMOS LIII: Passions and Emotions. New York University Press, 19–37.
Frazer, Michael, 2015. ‘Anatomist and painter: Hume’s struggles as a sentimental
stylist’, in Heather Kerr, David Lemmings and Robert Phiddian, eds., Passions,
Sympathy and Print Culture: Public Opinion and Emotional Authenticity in
Eighteenth-Century Britain. New York: Palgrave Macmillan.
Frazer, Michael, Forthcoming. ‘Interdisciplinary before the disciplines: sentiment-
alism and the science of man’, in R. Debes and Karsten Stueber, eds., Taking
Ethical Sentimentalism Seriously. Cambridge University Press.
Gewirth, Alan, 1968. ‘Meta-ethics and normative ethics’, Mind 69, 187–205.
Gewirth, Alan, 1970. ‘Metaethics and moral neutrality’, Ethics 78, 214–25.
Gibbard, Allan, 1990. Wise Choices, Apt Feelings: A Theory of Normative Judgment.
Cambridge, MA: Harvard University Press.
Goffman, Alice, 2014. On the Run: Fugitive Life in an American City. Chicago, IL:
University of Chicago Press.
Gross, Neil, 2013. Why Are Professors Liberal and Why Do Conservatives Care?
Cambridge, MA: Harvard University Press.
Guenther, Lisa, 2013. Solitary Conﬁnement: Social Death and Its Afterlives.
Minneapolis: University of Minnesota Press.
Haidt, Jonathan, 2001. ‘The emotional dog and its rational tail: a social intuitionist
approach to moral judgment’, Psychological Review 108, 814–34.
Hall, Cheryl, 2005. The Trouble with Passion: Political Theory Beyond the Reign of
Reason. New York: Routledge.
Hume, David, 1985. Essays, Moral, Political, and Literary, ed. Eugene F. Miller.
Revised Edition. Indianapolis, IN: Liberty Fund.
110 Chapter 6

Hume, David, 2000a. A Treatise of Human Nature, eds. David Fate Norton and
Mary Norton. Oxford University Press.
Hume, David, 2000b. An Enquiry Concerning Human Understanding, ed.
Tom Beauchamp. Oxford University Press.
Kingston, Rebecca, 2011. Public Passion: Rethinking the Grounds for Political Justice.
Montreal: McGill-Queen’s University Press.
Krause, Sharon, 2008. Civil Passions: Moral Sentiment and Democratic Deliberation.
Princeton, NJ: Princeton University Press.
Marcus, George, Russell Neuman and Michael Mackuen, 2000. Affective
Intelligence and Political Judgment. University of Chicago Press.
Nichols, Shaun, 2004. Sentimental Rules: On the Natural Foundations of Moral
Judgment. Oxford University Press.
Nussbaum, Martha, 1990. Love’s Knowledge: Essays on Philosophy and Literature.
Oxford University Press.
Nussbaum, Martha, 1995. Poetic Justice. The Literary Imagination and Public Life.
Boston, MA: Beacon Press.
Prinz, Jesse, 2007. The Emotional Construction of Morals. Oxford University Press.
Prinz, Jesse, 2011. ‘Is empathy necessary for morality?’, in Amy Coplan and
Peter Goldie, eds., Empathy: Philosophical and Psychological Perspectives. Oxford
University Press, 211–29.
Rawls, John, 1996. Political Liberalism. Revised Paperback Edition. New York:
Columbia University Press.
Rawls, John, 1999. A Theory of Justice. Revised Edition. Cambridge, MA:
The Belknap Press of Harvard University Press.
Seneca, 2004. Letters from a Stoic: Epistulae Morales ad Lucilium, tr.
Robin Campbell. New York: Penguin Classics.
Singer, Peter, 1972. ‘Famine, afﬂuence, and morality’, Philosophy and Public Affairs
1, 229–43.
Slote, Michael, 2010. Moral Sentimentalism. Oxford University Press.
Smith, Adam, 1984. The Theory of Moral Sentiments, eds. A. L. Macﬁe and
D. D Raphael. Indianapolis, IN: Liberty Fund.
Smith, Adam, 1985. Lectures on Rhetoric and Belles Lettres, ed. J. C. Bryce.
Indianapolis, IN: Liberty Fund.
Solomon, R. C, 1970. ‘Normative and meta-ethics’, Philosophy and
Phenomenological Research 31, 97–107.
Stevenson, Charles, 1944. Ethics and Language. New Haven, CT: Yale University
Press.
Stewart, John, 2013. The Unity of Content and Form in Philosophical Writing:
The Perils of Conformity. New York: Bloomsbury.
Thacher, David, 2006. ‘The normative case study’, American Journal of Sociology
111, 1631–76.
Thomson, Judith Jarvis, 1971. ‘A defense of abortion’, Philosophy and Public Affairs
1, 47–66.
Vermeule, Blakey, 2010. Why Do We Care about Literary Characters? Baltimore,
MD: Johns Hopkins University Press.
Moral Sentimentalism 111

Walzer, Michael, 2004. Politics and Passion: Toward a More Egalitarian Liberalism.
New Haven, CT: Yale University Press.
Weber, Max, 1978. Economy and Society, eds. Guenther Roth and Claus Wittich.
Berkeley: University of California Press.
Westen, Drew, 2007. The Political Brain: The Role of Emotion in Deciding the Fate of
the Nation. New York: Public Affairs.
Williams, Bernard, 1981. Moral Luck. Cambridge University Press.
7 Realism

Robert Jubb

1 Introduction
Contemporary normative analytical political theory tends to think of itself
as continuous with or at least an application of moral philosophy. For
example, in the Introduction to his Political Philosophy: A Beginner’s Guide
for Students and Politicians, Adam Swift describes the discipline he is
introducing as asking ‘what the state should do’, explaining that this
means asking ‘what moral principles should govern the way it treats its
citizens’ (Swift 2014: 5). Similarly, in the Introduction to his
Contemporary Political Philosophy, Will Kymlicka says that ‘there is
a fundamental continuity between moral and political philosophy’
(Kymlicka 2002: 5). This is because Robert Nozick was right to claim
that moral philosophy sets the limits to what ‘persons may and may not do
to one another’ including ‘through the apparatus of a state’ (Kymlicka
2002: 5). Nor is this only a feature of purportedly introductory texts.
Cecile Fabre’s Cosmopolitan War sees its attempt to articulate to integrate
cosmopolitan principles of distributive justice and just war theory as
uncomplicatedly an enquiry in ‘applied ethics’ (Fabre 2012: 3).
Equally, the very first sentence of Thomas Christiano’s The Constitution
of Equality: Democratic Authority and Its Limits introduces its investigation
into democratic authority by asking what ‘the moral foundations of
democracy and liberal rights’ are (Christiano 2008: 1).
Realism rejects this understanding of how normative political theory
should operate. This ‘moralist’ or ‘ethics first’ approach to normative
political theory is, realists claim, fundamentally mistaken. The precise
description of moralism’s sins varies from realist theorist to realist theor-
ist, but as realism’s reaction against moralism grows in strength, various
themes have emerged. This chapter treats Bernard Williams’s ‘Realism
and moralism in political theory’ as an archetypical piece of realism, using
it to suggest that realists share a common commitment to the idea that
politics is the contextually specific management of conflicts generated by
our inability to order our lives together around an agreed set of complete

112
Realism 113

moral values. It then goes on to try to demonstrate how that understand-

ing of politics constrains normative political theorizing. Realists have not
typically been very eager to move beyond critiques of moralism by enga-
ging in ﬁrst-order theorizing themselves. Although I discuss some reasons
why this may be, I nonetheless explore how realism might structure our
thinking about egalitarian political commitments. In the course of doing
so, I provide a series of guidelines a piece of political theorizing should
follow if it is to remain realist.
Section 2 discusses Williams and his realism. Section 3 provides a set of
guidelines for realist political theorizing by considering a realist case for
political egalitarianism. Section 4 summarizes the guidelines from the
previous section and includes a warning about the possibility of genuinely
political theorizing.

2 Bernard Williams’s Exemplary Realism

The current realist movement in political theory seems to have begun
to take off with the posthumous publication of a series of papers by the
British philosopher Bernard Williams. His collection, In the Beginning
Was the Deed, and particularly its first paper, ‘Realism and moralism in
political theory’ (hereafter RMPT), gave new and powerful voice to an
often long-held dissatisfaction with the dominant forms of political
theory in the Anglophone world (Williams 2005a, 2005b). Although
in many ways, Williams was there merely reiterating concerns he had
previously publicly aired, often in more polished forms, RMPT served
as a focus around which a range of complaints could coalesce. It has, for
example, around five times the number of citations as his last major
article, which, although it does not use the terms ‘realism’ and ‘moral-
ism’, is in effect an attempt to develop a realist theory of liberty by
taking into account various political constraints (Williams 2005c)1.
According to William Galston’s influential survey article, we even
owe Williams the term ‘realism’ as a way of grouping together those
dissatisfied with the way the ‘high liberalism’ of Rawls and Dworkin
ignores the centrality of conflict and instability to political questions
(Galston 2010: 386, 385). If we are to group together theorists as
different as John Dunn and Bonnie Honig, Chantal Mouffe and Mark
Philp, then it is to Williams we must look. It is then with Williams and
RMPT that I will start.

1
See https://scholar.google.co.uk/scholar?cites=10056158126568575885&as_sdt=2005&
sciodt=0,5&hl=en and https://scholar.google.co.uk/scholar?cites=395693232102231110
5&as_sdt=2005&sciodt=0,5&hl=en.
114 Chapter 7

RMPT begins with a distinction between two different ways of thinking

about the relation of ‘morality to political practice’ (RMPT: 1). While an
enactment model of that relation surveys society ‘to see how it may be
made better’ and so makes politics ‘the instrument of the moral’,
a structural model is instead concerned with ‘the moral conditions of co-
existence under power’ and so emphasizes ‘constraints . . . on what poli-
tics can rightfully do’ (RMPT: 1–2). While utilitarianism is
a paradigmatic case of the former, Rawls’s theory exemplifies the latter
(RMPT: 1). Despite the important differences that Williams identifies
between the enactment and the structural model, they share
a commitment to ‘the priority of the moral over the political’ and so
make political theory ‘something like applied morality’ (RMPT: 2).
Williams goes on to contrast this ‘political moralism’ with ‘political rea-
lism’, ‘which gives a greater autonomy to distinctively political thought’
(RMPT: 2–3).
Giving greater autonomy to distinctively political thought involves
focusing on what Williams calls ‘the first political question’ of ‘securing
of order, protection, safety, trust and the conditions of cooperation’
(RMPT: 3). This Hobbesian question is first not in the sense that it can
be solved and then ignored, but instead in the sense that its being and
remaining solved is a condition of posing, never mind solving, any other
political questions (RMPT: 3). For Williams, groups that do not attempt
to answer the first political question for themselves and their members do
not have politics. If order is not being created out of division in a way that
in some sense at least hopes to avert recognizably Hobbesian bads,
politics is simply not going on. Similarly, an activity or system of thought
is only political if it is circumscribed by a need to contain conflict among
those at whom it is aimed. As Williams puts it in the context of
a discussion of the relation between the Spartans and the Helots, the
‘situation of one lot of people terrorizing another lot of people is not per se
a political situation; it is rather the situation which the existence of the
political is in the first place supposed to alleviate’ (RMPT: 5). However
we want to characterize the undesirability of Hobbesian bads – and there
will surely be a range of ways to describe what is wrong with them –
politics is what goes on when we seek to avert them through coercive
orders.
Williams goes on to draw various further consequences from this
understanding of politics. He offers the idea of the basic legitimation
demand, which requires that some justification of claims to political
authority is offered to all those who are subject to it (RMPT: 4). After
all, the situation of the Spartans and the Helots is just one of any number
of examples that demonstrate that political power can contribute to the
Realism 115

problem of our absolute vulnerability to violence rather than offer

a solution to it. As Williams puts it, ‘something has to be said to
explain . . . what the difference is between the solution and the problem’,
to explain why the exercise of political power should not be treated simply
as an illegitimate attempt at domination (RMPT: 5). The basic legitima-
tion demand is the demand for that something. If it is a moral demand, ‘it
does not represent a morality which is prior to politics’ (RMPT: 5).
Instead, the basic legitimation demand follows from ‘there being such
a thing as politics’, from aiming to answer the first political question, from
showing that attempted solutions are not in fact ‘part of the problem’
(RMPT: 5).
Williams further claims that under contemporary conditions, only
liberalism can adequately answer the basic legitimation demand (RMPT:
7–8). This has not always been the case, but liberals have managed to
raise ‘expectations of what a state can do’, adopted ‘more demanding
standards of what counts as a threat to people’s vital interests’, and
expanded the range of ways in which supposed justifications can come
to seem like mere rationalizations (RMPT: 7). Answers to the basic
legitimation demand must always be historically variable in this sense:
they must ‘make sense’ of what they legitimate as ‘an intelligible order of
authority’ to those to whom it must be legitimated (RMPT: 10).
The universalist tendencies of political moralism that invite us to imagine
ourselves ‘as Kant at the court of King Arthur’ may offer us a genuine
possibility, but it is not a productive one (RMPT: 10). Performing that
thought experiment will not help us ‘to understand anything’ about
societies distant from ours in time and space, except perhaps that they
are distant not only in time and space (RMPT: 10). It is only when we
think about our own society and those it interacts with that what makes
sense as an intelligible order of authority becomes normative rather than
a ‘category of historical understanding’ (RMPT: 11).
Political moralism fails in terms of this understanding of politics.
By starting with a moralized and philosophical conception of the person,
it puts itself in a position from which it is impossible to give adequate
normative guidance about when Hobbes’s first political question has been
answered. A liberal conception of the person is the product and not the
justificatory foundation of the liberal political institutions we live under
now. Treating it as the foundation makes it impossible to provide an
explanation of those institutions or why we have them now and others,
elsewhere or at other times, do or did not (RMPT: 8–9). Further, because
political moralism sees politics as the application of moral philosophy to
political problems, it does not understand how to deal with political
disagreement. It ‘naturally construes conflictual political thought in
116 Chapter 7

society in terms of rival elaborations of a moral text’ and so sees its

opponents as ‘simply mistaken’ instead of fellow democratic political
actors whose deeply held commitments are at stake in political decisions
(RMPT: 12, 13).
Realism in general shares both RMPT’s hostility to much contem-
porary political philosophy and its diagnosis of its problems.
The survey articles by Galston and by Rossi and Sleat offer four
features characteristic of realism’s rejection of what it sees as the
moralism dominant in contemporary political philosophy (Galston
2010: 408; Rossi and Sleat 2014: 691–4). Although their lists do
not overlap perfectly, there is an understandable degree of similarity.
For example, Galston begins his list by claiming that realism involves
taking ‘politics seriously as a particular field of human endeavour’
and that this means holding that ‘civil order is the sine qua non for
every other political good’ (Galston 2010: 408). In turn, Rossi and
Sleat first stress the importance for realists of the ‘broadly Hobbesian
thought . . . that if ethics could effectively regulate behaviour in
political communities as it does amongst (say) friends and acquain-
tances, we would not require politics’ (Rossi and Sleat 2014: 691).
In both cases, this mandates an attention to ‘the specific conditions
under which political decisions are taken and agents act’ (Rossi and
Sleat 2014: 694), whether that be in terms of a focus on institutions
and a more developed moral psychology or on the history of our
moral commitments and the tragic choices that political actors may
find it impossible to avoid (Galston 2010: 408; Rossi and Sleat 2014:
691–4).
Many of the elements in terms of which Galston and Rossi and
Sleat define realism are present in RMPT. RMPT is hostile to much
contemporary political philosophy and theory on the grounds that it
is not properly political in one sense or another. Williams’s discus-
sion understands politics in terms of the provision of order for agents
whose interests and ideals conflict in a way that otherwise might well
make it impossible for them to coexist. In stressing the importance of
conflict to political thinking, Williams here also insists on the impor-
tance of context. The universalist tendencies of contemporary poli-
tical philosophy and theory are part of what prevent it from
addressing real political situations, which always involve actual poli-
tical actors with particular disagreements. Those disagreements and
the resources the situation makes available to resolve them need to
be properly understood if anything helpful is going to be said about
them. This will mean appreciating how we came to find ourselves
here, with these conflicts and these means of defusing and
Realism 117

controlling them. The history and speciﬁcity of our situation need to

be understood so that we can grasp the limits on what we can do.
Although Williams expresses scepticism about what he sees as
Habermas’s project to show that ‘the concept of modern law har-
bours the democratic ideal’ in RMPT (16), he acknowledges that the
discussion in RMPT occurs at ‘a very high level of generality’ and so
does not contain any concrete positive claims of the sort he criticizes
Habermas for making (15). In that sense, despite being the best
known of Williams’s realist pieces, there is not really any ﬁrm advice
about how to do realist political theory or philosophy in RMPT.
Even elsewhere, when for example Williams is discussing the parti-
cular political value of liberty, although he is eager to tell us how not
to judge whether someone really does have a complaint in liberty,
there is little positive theorizing. We are told that competition is not,
for us here and now, the ground of a complaint that liberty has been
lost but that one’s position in a social structure can be, and that
because of our disenchantment, liberty is more important to us than
many of our forebears, all at roughly this level of generality (Williams
2005c: 91, 95). Nor is Williams unusual here. Self-identiﬁed realists
have been much more interested in diagnosing problems with con-
temporary political philosophy than replacing the positive theorizing
they criticize.
Part of the reluctance to be more forthcoming here is undoubtedly
the importance Williams and other realists give to political action. Even
if the title were chosen by Williams’s editor and widow, it is obviously
no accident that the posthumous collection containing RMPT is called
In the Beginning Was the Deed. Williams treats that dictum from
Goethe’s Faust as a reminder that politics is about action and so will
often escape our attempts to model or predict it because of the way its
participants’ acts will transform it, including by creating the conditions
of their own success. As well as quoting it in RMPT, Williams uses
Goethe’s dictum as the title for another of the pieces in the posthumous
collection (RMPT: 14; Williams 2005d). However, unless realists think
that political theory or philosophy is a necessarily impossible activity, it
must be capable of at least sometimes meeting those conditions.
Indeed, Williams’s own career, which involved sitting on a number of
Royal Commissions and contributing to the British Labour Party’s
Commission on Social Justice, suggests that he felt that political theory
and philosophy could address concrete political questions without fall-
ing victim to the pathologies of moralism. By in part drawing on some
of Williams’s own work on equality, the remainder of this piece will try
to illustrate how that might be done.
118 Chapter 7

3 Working through a Case: Legitimacy and a Realistic

Egalitarianism
Williams insisted that only a liberal state could be legitimate under the
conditions of modernity (see for example RMPT: 7–8). It is this, and the
arguments this might give us for commitments to relatively high levels of
material equality, that will serve as examples to demonstrate how to do
positive realist political theorizing. The first task here is to understand
why Williams thought that liberalism was the only way of legitimating
a state in modernity. Although political situations must involve the man-
agement of conflict between agents whose commitments and interests
cannot all be satisfied, for realists they are never exhausted by that
characterization. Responding to a political situation then will have to
mean responding to its particularities. Realists must rely on an inter-
pretation of a political situation that captures its specificities;
otherwise, they will be guilty of the universalism and the associated failure
to address real political agents for which they criticize moralists.
Williams believed that modernity required liberalism because it had
raised expectations of what the state could do while undermining the ease
with which hierarchies can be justified (see for example RMPT: 7).
The idea that modernity involves the triumph of rationality over mystical
and supernatural explanations is a persistent theme throughout
Williams’s work, from ‘The idea of equality’ (hereafter IoE) in 1962 to
Truth and Truthfulness in 2002 (Williams 2005e: 105; 2002: 231).
If hierarchies of rank of the sort liberals tend to reject are not to depend
on brute coercion, they must rely on seeming ‘foreordained and inevita-
ble’ and so are ‘undermined’ by growth of their members’ ‘reflective
consciousness’, especially about the way that such hierarchies tend to
enculturate their members (IoE: 105). If a modern political order was to
be justified then, for Williams, it had to be justified to Weberian disen-
chanted agents. A realist theory of a political good like legitimacy
must be fitted to the particular political situation in which it is to
be invoked.2
Of course, Williams’s interpretation of our situation now and around
here is hardly uncontested. One might think of Alasdair MacIntyre’s
insistence that something roughly like the processes that undermined
Williams’s ‘supposedly contented hierarchical societies of the past’ were
a disaster for reflective moral understanding analogous to the destruction
of science as a practice of investigation and understanding (MacIntyre

2
There is in this sense a link between practice-dependence and realism. See Sangiovanni
(2008) for a deﬁnition of practice-dependence and Jubb (2016) for discussion of the
relation between the two.
Realism 119

1981: 1–2; Williams 2011: 181). If modernity has left us with ‘fragments
of a conceptual scheme’ stripped of the ‘contexts from which their sig-
nificance derived’ in the place where integrated notions of the good life
ought to be, then Williams’s support for liberalism will seem, at best,
acquiescence in a cultural catastrophe of an unimaginably vast scale
(MacIntyre 1981: 2, 3). MacIntyre’s stance here is at least in tension
with realism because of the way in which it refuses to deal with the agents
with which it understands itself as being faced. For MacIntyre, we are
doomed by the complete disintegration of the traditional authorities that
made possible the Thomist virtues we need for decent lives. How does
that diagnosis of our problems tell us to structure our lives together?
It must absolutely reject not just those institutions and their associated
historical and sociological forms but with them, us.
In this sense, a realist political theory must be based on an inter-
pretation of our political situation that refuses both the related
consolations of utopian hope and unremitting despair. MacIntyre
believes that modernity makes it impossible for us to live decent lives. He
combines utopianism with despair by claiming that unless we undo all the
history of at least the past three centuries, we are doomed to live fractured,
empty lives. The rejection of everything there is prompts the search for
something beyond it. This is not to say that such interpretations or even
the commitments for which they serve as foundations are incorrect or
inappropriate. It is instead to point out that if any really achievable social
order destroys all but the most minimal human values, then it is hard to
understand the point or even the possibility of ‘securing order, protection,
safety, trust, and the conditions of cooperation’ (RMPT: 3). There will be
no political values for distinctively political thought to articulate. Whether
they are conservative or radical, realists must be able to say something that
can make sense of the value of politics as an activity. The temptation to
slip into moralist condemnation needs to be resisted.
Emphasizing this may cast some doubt on the credentials of some self-
proclaimed realists, at least if they intend their realism to involve ‘distinc-
tively political thought’ in Williams’s sense (RMPT: 3). Raymond Geuss,
for example, not only sees himself operating in the tradition of critical
theory typified by Theodor Adorno; he criticizes Williams for ‘paddling
about in the tepid and slimy puddle created by Locke, J. S. Mill and Isaiah
Berlin’ rather than adopting Adorno’s rejection of both the ‘self-serving
“liberalism” of the Anglo-American political world and the brutal prac-
tices of “really-existing socialism”’ (Geuss 2012: 150). He also sees
himself as a realist; his Philosophy and Real Politics is a relentless attack
on behalf of realism against what he calls ‘ethics-first’ political philosophy
(see for example Geuss 2008: 9). However, anyone taking their lead from
120 Chapter 7

Adorno may find themselves too pessimistic about our historical and
political situation to be able to do justice to what we can achieve through
politics.
Adorno argued that while ‘social freedom is inseparable from
enlightened thought,’ ‘[t]he only kind of thinking that is sufficiently
hard to shatter myths is ultimately self-destructive’ (Adorno and
Horkheimer 1979: xiii, 4). If as a result our civilization is in fact
a kind of ‘barbarism’, its social orders will presumably be little more
than ‘one lot of people terrorizing another lot of people’ in more and
less open ways (Adorno and Horkheimer 1979: xi; RMPT, 3). That
will make it difficult to understand how, for example, Williams’s first
political question could be answered satisfactorily. Geuss may then
not be a realist, at least in the sense I am using here. Following
Adorno, he insists both that living a decent life is impossible within
the ‘repressive, duplicitous and alienated’ social forms of late capitalist
modernity and that it is impossible for us realistically ‘to envisage any
fundamental change in our world we could bring about by our own
efforts’ (Geuss 2012: 154, 160). If our historical and political situation
means that our lives cannot avoid being ‘radically defective’, an
‘impossible situation’, then we will struggle to find, let alone imple-
ment or grasp the value of, ways of living together that do not betray
all our hopes (Geuss 2012: 154).
There are ways to reject Williams’s account of modernity that do not
see it as made up of ideas that are ‘misshapen, brittle, riven with cracks . . .
and very ill-suited to each other’, and so under which it is impossible to
live coherently (Geuss 2001: 9). We might, for example, question
whether Williams can be right about liberalism being the only way that
a modern political order can make sense to its members. One does not
have to subscribe to claims about the superiority of alleged Asian values to
see that various states in East Asia seem to be accepted by most of their
citizens yet are neither liberal nor underdeveloped compared to the North
Atlantic democracies Williams presumably had in mind when equating
liberalism and modernity. There are problems too for Williams’s claim
even in Europe, where we might assume it would be most apt given its
association with the Enlightenment and its supposedly demystifying
aftermath. Many European states, most obviously those of the former
Warsaw Pact, have been and in some cases remain modern and illiberal
without obviously failing to give a broadly acceptable account of them-
selves to their citizens, even in the medium term. Even if modernity is
disenchanted, it seems that there are a variety of ways of responding to
that disenchantment. The interpretation of the relevant political
situation on which a realist relies must not generate obviously
Realism 121

implausible implications, as Bernard Williams’s does if he is taken to

be discussing modernity in general.
Rawls is often the target of realists’ attack on moralism. He exemplifies
one of Williams’s two forms of moralism, while Galston begins his survey
of realism by rightly describing it as a ‘countermovement’ to the ‘high
liberalism’ championed by Rawls and Dworkin. However, in his later
works, Rawls drew a distinction between political and comprehensive
theories and defended the political credentials of his own work (Rawls
2005). Part of this involved situating his theorizing ‘in the special nature
of democratic political culture as marked by reasonable pluralism’ (Rawls
2005: xxi). That culture has its roots in ‘the doctrine of free faith’ devel-
oped in the aftermath of the Reformation that rejects the idea that ‘social
unity and concord requires agreement on a general and comprehensive
religious, philosophical, or moral doctrine’ that might otherwise have
seemed natural in a world of ‘salvationist, creedal, and expansionist
religions’ (Rawls 2005: xxv). These conditions do not characterize all
twentieth-century European societies, let alone all societies that might be
described as modern, as for example Rawls’s references to the Weimar
Republic’s loss of confidence in a ‘decent liberal parliamentary regime’
show he was well aware (Rawls 2005: lix). In that sense, not only do Rawls
and realists share concerns about treating political philosophy as a branch
of ethics, but Rawls is in fact clearer than Williams is about the relatively
restricted scope of the interpretation of the political situation on which his
principles rely.
Realist disagreement with Rawls should not focus on his alleged failure
to address properly political questions, since his later theory is in fact
explicitly arranged around addressing a particular, historically situated
form of disagreement, but on the tools with which he chooses to address
them. The problem with Rawls’s theory is not that it ignores Hobbes’s
question, but that it treats it as soluble through appeals to an ideal of free
and equal citizenship (see for example Rawls 2005: xxv). Rawls is wrong
to think that philosophical abstraction can by itself offer a way of dealing
with ‘deep political conflicts’ like those between Lincoln and his oppo-
nents over slavery, because those conflicts are obviously not only philo-
sophical disagreements (Rawls 2005: 44, 45). Rawls’s understanding of
the dilemmas of modern democratic life is inadequate because it ignores
both the role of material interests in our political life and the cognitive and
motivational limits of philosophical reasoning. When Rawls developed his
theory, in the long period of Keynesian growth after World War II, there
were real political movements that officially had plausible hopes for some-
thing like what he prescribed. Although in that sense, his views are not
pejoratively utopian, his inadequate understanding of the dilemmas of
122 Chapter 7

modern democratic life means they are not realist. Those political move-
ments did not draw on philosophical ideals to draw together and motivate
their supporters, but on shared experiences of hardship and solidarity
built up in the course of struggles against it. Realists must acknowledge
the importance of material interests and ideological and charis-
matic appeals, especially compared to philosophical reasoning,
when theorizing political goods to ﬁt particular political
situations.
Rawls’s and Williams’s interpretations of modernity illustrate two
errors to which interpretations of a political situation may fall victim.
Complacency about the generality of an interpretation or about the
motivational and cognitive power of philosophy is not realistic.
However, we can avoid both of those problems by marrying Rawls’s
cultural and geographical circumspection to Williams’s emphasis on dis-
enchantment. Such an interpretation of modernity will neither apply
beyond the societies with which both theorists were familiar nor end up
depending on the power of reason alone. This leaves us with a roughly
liberal principle of transparency, which, rather than requiring a system
capable of being endorsed by all ‘in light of principles and ideals accep-
table to their common human reason’, builds on Williams’s idea of the
‘human point of view’ (Rawls 2005: 137; IoE, 103).
The human point of view is a perspective that, when considering some-
one’s else’s life, ‘is concerned primarily with what it is for that person to live
that life and do those actions’ (IoE: 103). It asks us to ‘respect and try to
understand other people’s consciousness of their own activities’, whether
they manage to do what they hoped and how they feel about failures they
suffer (IoE: 103). The human point of view can become a demand for
transparency and rule out, Williams claims, markedly unequal societies
because of the frustrations and resentments they will predictably gener-
ate. Once the idea that societies are human creations spreads, those
frustrations and resentments can no longer be justiﬁed because they will
no longer seem inevitable (IoE, 105). If relations between members of
those societies are not to be conducted on the basis of brute force or
systematic deception, neither of which can be acceptable from the human
point of view, then hierarchies cannot be too steep or pervasive (IoE:
104–5). If the hierarchies are too steep or pervasive, then the social and
political order that sustains them will seem too unsympathetic to the
‘intentions and purposes’ of those at the wrong end of those hierarchies
(IoE: 103). The terms on which that sympathy operates will need to be
thinner, to assume less about the commitments of those whom it is for,
than the terms on which it operated in ‘supposedly contented hierarchical
societies of the past’ (see for example Williams 2005f). It will be unable to
Realism 123

take for granted the set of commitments we all supposedly shared in those
societies. Those are gone, along with metaphysical or supernatural
explanations that sustained the idea that the associated hierarchies were
unavoidable. Still, that sympathy will need to be there if the political order
is to ‘make sense’ to its members, to avoid failing to meet the basic
legitimation demand for at least those particularly disadvantaged by it.
In understanding what this account of legitimacy in North Atlantic
democracies might judge acceptable, we should look to real political
motivations. These need not be drawn directly from reality, as long as it
is clear that they have some real-world counterparts (see for example Jubb
2015a). Our accounts of political legitimacy or any other political
good must be for actually existing agents, and we can best check
that there is a constituency that they address by showing that they
can capture and give form to political demands that animate
actual agents. If we were to develop an account of a political good that
could not be seen as an articulation of a hope or resentment that drives
a stance towards a political order real people actually adopt, then that
failure would count strongly against the account being genuinely realist.
For example, connecting an interpretation of Williams’s minimally ega-
litarian account of legitimacy to the resentments that seem to have moti-
vated the most widespread civil unrest in the United Kingdom in recent
decades strengthens that account by showing that it could well make
sense of those real political demands (see Jubb 2015b). A series of inter-
views conducted with hundreds of self-identiﬁed rioters found
a ‘pervasive sense of injustice’ (Lewis et al. 2011: 24). Barely half of the
rioters felt British, compared to more than 90 per cent of Britons on
average, understandably given that they felt victimized by the police and
excluded from a culture of consumption by their poverty (Lewis et al.
2011: 28, 19). These interviews seem to show then that a realist egalitar-
ianism focusing on the systematic frustration of the hopes and expecta-
tions of the least advantaged speaks to real political motivations.
Realism does not just demand that political goods are for actual agents.
Actual agents may of course make demands that are not properly political
and so realists will have to temper and limit them. Indeed, the intense
moralism of much democratic political debate, which is for example often
captivated both on the left and the right by nostalgia for supposedly lost
forms of ethical community, may be a serious problem for realists. Even if
that moralism can be contained within the boundaries of the properly
political for now, there is presumably always a risk that dissatisfaction
with the inevitable compromises of political life will break through those
limits and put various political goods at risk. Descriptions of political
goods meant for moralistic publics will have to explain in terms
124 Chapter 7

they can understand why they must satisfy themselves with less
moral unity than they would like. In this sense, realist political theory
needs to draw not just on an interpretation of a political situation, but also
on an interpretation that is capable of being publicly stated and accepted
without, for example, undermining itself. If an account of legitimacy
shows a population they share less than they thought, it may prevent
them from sharing even that.
This is just one of a number of risks that realist pieces of theorizing face
as a result of the constraints imposed by the need to remain political.
The most obvious of these is that the theorizing itself relies on contro-
versial moral values. Realist political theorizing would be moralistic in this
sense if it relied on value claims that, if they were acceptable to the
constituency the theory addresses, would eliminate its political problems.
Politics is in part constituted by our disagreement on values around which
to order our shared institutions. Consequently, realist political theory
must not appeal to values or interpretations of values whose con-
troversy is, at least as far as its interpretation of the relevant
political situation is concerned, a deﬁning feature of that
situation.
For example, if we were to try to articulate an egalitarian theory of
legitimacy along the lines suggested by Williams’s idea of the human
point of view, we would need to avoid basing its appeal on an ideal of the
good of living as equals. While perhaps there have been some commu-
nities where such an ideal could exert enough power over most of its
members to distinguish between legitimate and illegitimate rule, our
society is not one. It is too divided to demonstrate the acceptability of its
massively coercive, structuring power by appealing to the idea that
relations of equality are central to a good life. Even if that ideal seems
attractive in the abstract, its acceptability in general will not decide how
it should be weighed against other ideals when it inevitably conﬂicts with
them. A solidaristic society may be all well and good for many citizens,
but only as long as it does not suppress individuality, undermine indi-
vidual responsibility, respect for one’s traditions or any number of other
ideals.
Instead, an egalitarian theory of legitimacy must be based on a less
demanding account of the value of equality. Rather than requiring
citizens to accept not just a moral ideal, but a particular ranking of that
ideal against other competing ideals, some more minimal explanation
of the value of equality and its connection to an entitlement to rule is
needed. For Williams, one of the virtues of Judith Shklar’s liberalism
of fear was that it tried to address ‘everybody’ by drawing on ‘the only
certainly universal materials of politics . . . power, powerlessness, fear,
Realism 125

cruelty’ (Williams 2005g: 59). As Shklar herself understood, the lib-

eralism of fear’s emphasis on minimizing our exposure to cruelty
could be turned against hierarchy because of the way that abuses of
power ‘are apt to burden the poor and weak most heavily’ (Shklar
1989: 28). Inequality often brings domination and humiliation in its
wake, and so the importance of equality could be explained by trying
to avoid those harms. Domination and humiliation count as harms in
terms of many plausible ideals, and so understanding the value of
equality through the value of avoiding them would minimize conﬂict
between that value and others. Since generating and sustaining dom-
ination and humiliation seems to make power relations illegitimate, it
would also connect answering the basic legitimacy demand with meet-
ing various egalitarian requirements. The values to which a realist
political theory appeals must be minimal in the sense that they
can expect to be accepted as playing whatever role is necessary
by at least most members of the society to which they must
‘make sense’.
It is not enough to be able to say that, for example, inequality causes
domination and humiliation. Insofar as an explanation of what
matters about a particular value depends on claims about how
it links to other values, those claims have to be substantiated by
an empirically sensitive account of how social and political life
actually operates. For example, G. A. Cohen claims that market
interactions are ‘typically’ motivated by ‘some mixture of greed and
fear’ in that other participants in markets ‘are predominantly seen as
possible sources of enrichment, and threats to one’s success’ (Cohen
2009: 40). This is supposed to contrast with and so help explain the
value of an alternative motivation of community, which values recipro-
cal service (Cohen 2009: 39–45). It is, though, straightforwardly false
that market interactions are predominantly structured around greed
and fear. The norms of basic honesty and respect for property rights on
which a functioning market depends could not survive if we all saw each
other primarily as ruthless exploiters desperately hiding our vulnerabil-
ities from each other to avoid them being taken advantage of. Cohen’s
account of the value of community is discredited by the obviously
inadequate picture of human interactions on which it partly depends.
In contrast, linking equality with avoiding domination and humiliation
seems to have some empirical support. Work in social epidemiology
like that of Michael Marmot, Richard Wilkinson and Kate Pickett
suggests that inequalities tend to ossify into status hierarchies that
dominate and humiliate those at the wrong end of them (Marmot
2004; Wilkinson and Pickett 2009).
126 Chapter 7

4 A Summary and a Warning

At this point, we can summarize the guidelines for which I have so far
argued and which together seem to me deﬁne realism, at least when
contrasted to moralism in something like Bernard Williams’s sense.
1. A realist account of a political value must be based on an inter-
pretation of the political situation in which the value is to be
realized.
2. That interpretation of the situation must be plausible, not least
in avoiding both relentless despair and utopian hope.
3. The value being theorized must be one that agents can be
expected to respect as the theory requires without becoming
moral saints.
4. Actual agents should also be able to see something of their
expectations or aspirations in the theory that is being offered
for their political situation, even if they may have more expan-
sive hopes than it makes room for.
5. That theory should not rely on controversial interpretations or
rankings of values, but try to make use of the evaluative and
normative material the situation presents.
6. When connections are drawn between different normative or
evaluative claims, as they will have to be, these connections
must be based on plausible theories of and claims about how
human life actually operates.
These guidelines distinguish realism from moralism. Moralism is uni-
versalist and uninterested in the details of particular situations. Swift,
Kymlicka, Fabre and Christiano are all engaged in ahistorical projects of
justiﬁcation in the works from which I quoted in the Introduction to this
chapter. Nor do they see it as problematic to criticize individuals or the
world for failing to live up to ideals in which they or it obviously have little
or no interest. Few actually existing states meet or look likely to meet
Christiano’s demanding criteria for democratic legitimacy, for example
(see e.g. Christiano 2008: 260–1). Certainly his own polity, the United
States of America, is a very long way from meeting those criteria. Nor is it
clear that those criteria, or those that Swift, Kymlicka and Fabre endorse
in their books, relate to real political aspirations held by those outside the
academy. Finally, the four do not seem to feel a need to show that the
values they theorize can be integrated into a realistic picture of how
human social and political life actually operates.
Following the directions I have given should make a realist theory of
a particular political value adequate for a particular situation at
a comparatively general level. There will be nothing about the theory
Realism 127

itself that prevents it from making sense of the value of whichever political
projects it favours to those for whom it favours them. Still, being in
principle able to articulate a given political value to and for a particular
group of people does not mean that the articulation will satisfy or be
accepted by those people. Nor does it mean that they will actually be able
to organize themselves into a collective capable of achieving whatever it
demands or hopes. To move beyond generic and towards what we could
call full realism, a theory should not just eliminate barriers to providing an
account of a particular political value to those in a particular political
situation. A realist political theory should also show that its poli-
tical projects can capture and hold the allegiance of people against
the rival political projects that are bound to challenge them, and
that the supporters who can be attracted can collectively put them
into practice.
This will mean understanding the political, social and economic
dynamics operating in particular societies. For example, an egalitarian
realist theory of legitimacy seems to face at least three related questions
raised by the requirement that it show not only how it is generically, but
also fully, concretely realist (see Jubb 2015b for more detail on these
challenges for a realist egalitarianism). If a realist theory of legitimacy is to
sensibly demand that states restrict their levels of inequality or risk
becoming illegitimate, it needs to show, first, that enforcing limits on
inequality will not, as a matter of fact, undermine various other values.
After all, for example, equality is often associated with societies and
groups that tend to repress difference, and so we might find that however
desirable equality is in theory, in practice there is no way of achieving it
that does not compromise too many of our other commitments. Second,
it must be politically possible to fulfil the demands of an egalitarian realist
theory of legitimacy. If supporters of higher levels of equality cannot
dominate the political scene, or if their dominance would inevitably
bring about economic collapse caused by, say, capital flight, then we
will have to change our attitudes. Either we will have to understand
ourselves differently, as needing a different kind of explanation of what
our states must do for us, or, alternatively, we will have to see our state as
an alien, dominating force for at least some of its members. Third, we
need to have a reasonable expectation that a constituency can be united
around the indignities of inequality, and that they will not seek to deal
with the frustrations associated with inequality in other ways. Otherwise,
it would not be clear how the theory answered a real rather than imagined
problem.
Political theorists are often unlikely to be able to meet these require-
ments. Indeed, even scholars with empirical expertise may not often be
128 Chapter 7

able to meet them given the unpredictability of political life. In this sense,
political realism’s emphasis on politics as a distinctive sphere of life limits
the role of scholars, especially given the importance of action in that
understanding of politics. The deliberately modest understanding of
political theorists as ‘democratic underlabourers’ offered by Adam Swift
and Stuart White, for example, seems in fact inappropriate and over-
ambitious (Swift and White 2008: 54). The problem is not primarily, as
Swift and White worry, that offering philosophical arguments will bypass
the proper democratic process (Swift and White 2008: 55). It is instead
that philosophical arguments are dangerously unsuited to political pro-
blems. Nor will positivist empirical theories of political processes often be
any better off. Politics shapes the problems with which it has to deal by
shaping the agents, both individual and collective, whose motivations and
dispositions create its problems. A theory of politics capable of under-
standing all the processes relevant to its own applicability would be too
complex for humans to understand, and of course itself a tool that, were it
understood, it would have to include in its assessment of the relevant
dynamics.
Political theories are in this sense necessarily incomplete. Politics is
a sphere of judgement instead of scientiﬁc understanding, which will be
vindicated by the acts it recommends having the intended effects and so
after the fact, once the situation has been changed. No political theory can
show completely that it captures and deals appropriately with a particular
political situation, and so no political theory can be fully realistic. Like
judgements, though, political theories can be better and worse. Realists
believe that working with the guidelines they provide will at least make
them more likely to avoid failing by not being about politics at all.

Acknowledgements
I sent earlier drafts of this to Ed Hall, Enzo Rossi, Paul Sagar, Matt
Sleat and Patrick Tomlin, all of whom were kind enough to send very
helpful comments. Adrian Blau’s editorial suggestions also substan-
tially improved the piece. I am grateful to all of them, and even more
so to Adrian for asking me to contribute to this volume in the ﬁrst
place.

References
Adorno, Theodor and Horkheimer, Max, 1979. The Dialectic of Enlightenment.
London: Verso.
Christiano, Thomas, 2008. The Constitution of Equality. Oxford University Press.
Realism 129

Cohen, G. A., 2009. Why Not Socialism? Princeton, NJ: Princeton University
Press.
Fabre, Cecile, 2012. Cosmopolitan War. Oxford University Press.
Galston, William, 2010. ‘Realism in political theory’, European Journal of Political
Theory 9: 385–411.
Geuss, Raymond, 2001. History and Illusion in Politics. Cambridge University
Press.
Geuss, Raymond, 2008. Philosophy and Real Politics. Princeton, NJ: Princeton
University Press.
Geuss, Raymond, 2012. ‘Did Williams do ethics?’, Arion 19: 141–62.
Jubb, Robert, 2015a. Playing Kant at the court of King Arthur’, Political Studies 63:
919–34.
Jubb, Robert, 2015b. ‘The real value of equality’, The Journal of Politics 77:
679–691.
Jubb, Robert, 2016. ‘“Recover it from the facts as we know them”: practice-
dependence’s predecessors’, Journal of Moral Philosophy 13(1): 77–99.
Kymlicka, Will, 2002. Contemporary Political Philosophy: An Introduction. Second
Edition. Oxford University Press.
Lewis, Paul, Tim Newburn, Matthew Taylor, Catriona Mcgillivray,
Aster Greenhill, Harold Frayman and Rob Proctor, 2011. Reading the Riots:
Investigating England’s Summer of Disorder. London: The London School of
Economics and Political Science and The Guardian.
MacIntyre, Alasdair. 1981 After Virtue: A Study in Moral Theory. Notre Dame, IN:
University of Notre Dame Press.
Marmot, Michael, 2004. Status Syndrome: How Your Social Standing Directly Affects
Your Health. London: Bloomsbury.
Rawls, John. 2005. Political Liberalism. 2nd edition. New York: Columbia
University Press.
Rossi, Enzo and Matt Sleat, 2014. ‘Realism in normative political theory’,
Philosophy Compass 9/10: 689–701.
Sangiovanni, Andrea, 2008. ‘Justice and the priority of politics to morality’, Journal
of Political Philosophy 16: 137–64.
Shklar, Judith. 1989. ‘The liberalism of fear’, in Nancy Rosenblum, ed., Liberalism
and the Moral Life. Cambridge, MA: Harvard University Press. 21–38.
Swift, Adam, 2014. Political Philosophy: A Beginner’s Guide for Students and
Politicians. Cambridge: Polity.
Swift, Adam and Stuart White, 2008. ‘Political theory, social science, and real
politics’, in David Leopold and Marc Stears, eds., Political Theory: Methods and
Approaches. Oxford University Press, 49–69.
Wilkinson, Richard and Kate Pickett, 2009. The Spirit Level. London: Allen Lane.
Williams, Bernard, 1997. ‘Forward to basics’, in J. Franklin, ed., Equality. London:
Institute for Public Policy Research, 47–57.
Williams, Bernard, 2002. Truth and Truthfulness. Princeton, NJ: Princeton
University Press.
Williams, Bernard, 2005a. In the Beginning Was the Deed. Princeton, NJ: Princeton
University Press.
130 Chapter 7

Williams, Bernard, 2005b. ‘Realism and moralism in political theory’, in Williams

2005a, 1–17.
Williams, Bernard, 2005c. ‘From freedom to liberty: the construction of a political
value’, in Williams 2005a, 75–96.
Williams, Bernard, 2005d. ‘In the beginning was the deed’, in Williams 2005a,
18–28.
Williams, Bernard, 2005e ‘The idea of equality’, in Williams 2005a, 97–114.
Williams, Bernard, 2005f. ‘Modernity and the substance of ethical life’, in
Williams 2005a, 40–51.
Williams, Bernard, 2005g. ‘The liberalism of fear’, in Williams 2005a, 52–61.
Williams, Bernard, 2011. Ethics and the Limits of Philosophy. London: Routledge.
8 Realistic Idealism

David Schmidtz

1 Theorizing for a Reason

Is moral philosophy more foundational than social philosophy? Is the
question of how to live more fundamental than the question of how to
live in a community? Are we getting down to philosophical foundations
when we set aside contingencies regarding the communities in which
political animals live, and proceed as if we were pure rational wills?
I see no reason to say yes to any of these questions.
To recover a measure of relevance to questions that practitioners need
to answer – questions about how to live as social beings – theorizing about
how to live together might take its cue less from moral philosophy and
more from political economy. We can go beyond thought experiments.
We can ask which principles have a history of being the organizing
principles of ﬂourishing communities. Let’s say that realism studies the
human condition as it is, while idealism studies the human condition as it
should be.1 Thus characterized, realism and idealism are distinct but
compatible projects.
Realistic idealism, one of the many possible forms of idealism, studies
what should be in light of a sober assessment of what could be, here and
now. It aims to identify real possibilities, then ask whether an ideal
response is among those possibilities. Realistically, it need not be.
An ideal response is a best response, and intuitively something more: we
call the best available response ideal only if we accept some fairly strong
version of the thought that we could not have done better. Suppose we say
Plan A is ideal, then ﬁnd that Plan A is no solution at all – maybe it is
infeasible because a key ingredient is missing. When we switch to our
actual best response, Plan B, we do so with regret about a solution that
seemed within reach and that would have been better. If we also restock the
missing ingredient so that Plan A will be a real option next time, that
implies that Plan B is merely best under the circumstances, not ideal.

1
See also Robert Jubb’s chapter on realism in this volume, and Sleat (2013).

131
132 Chapter 8

Realistically, not all problems have solutions, let alone ideal solutions.
Sometimes studying a problem helps us see what would solve it.
Sometimes we learn that the best we can do is mitigate. To introduce
the main bits of advice given in this chapter:
(a) Start with problems. We were taught to see sound theory as
grounding sound practice and therefore as needing to come ﬁrst.
In practice, theories are answers; questions come ﬁrst.
(b) Start with diversity. We need to coordinate on terms of engage-
ment that are apt even among people who do not agree that those
terms are apt. Theorizing does not help. We navigate the terrain
of respect for separate persons with a compass far older than any
theory.
(c) Start with injustice, not justice. In the real world, we have no
vision of ‘peak justice’ in mind when deciding how to act. When
deciding which car to try to drive home at the end of the day, we
never consult the theory of justice we spent all day perfecting, except
in self-mockery. Theory is not what teaches us how to avoid trigger-
ing people’s sense of injustice.
Finally, if I had a bit of meta-advice about how to handle my advice, it
would be: proceed with caution. Some of my advice will survive the test
of time and turn out to be good advice, but there is no substitute for
exercising your own judgement, being sceptical of contemporary theoriz-
ing about morality and justice, and taking your cue from the world rather
than from the literature.

2 We Are Political Animals

One enduring feature of the human condition is that we are, after all,
political animals. (1) We are decision-makers. (2) We are decision-
makers who want and need to live together. (3) As decision-makers, we
respond to circumstances. (4) As social beings, we respond to the circum-
stance that we live among decision-makers – other political animals who
treat our choices as part of their circumstances and respond accordingly.
Social theory done well is theory about a world of separate persons –
separate not only in an aspirational Kantian moral sense, but in
a straightforward descriptive sense that each person is a locus of agency.
People decide for themselves. We choose well only if we choose with
a view to what we thereby give others a reason to do in response – that is,
only if we do not take others for granted, do not treat them as pawns, and
do not treat them as if they have a duty to be gripped by whatever vision is
gripping us at the moment. If we are not theorizing along those lines, then
we are not theorizing about politics.
Realistic Idealism 133

Consider how the mundane observation that we are political animals

implies a need to take a slightly but importantly different approach to
moral theory (as developed in Schmidtz 2016; the following subsection is
an overture for the argument set out in that essay).

2.1 Solipsism in Theory

(a) Kantians regard ‘What can be universalized?’ as a foundational ques-
tion. People interpret that as a rough equivalent of ‘what if everyone did
that?’ The subtle but crucial piece missing from this informal rendering:
moral questions are questions for political animals living in a social world,
which means a strategic world. The actual problem moral agents face is
not a question of what maxim they could will everyone to follow.
In a strategic world, interpreting universalizability in solipsistic terms –
imagining a choice between everyone cooperating and everyone declining
to cooperate – is not universalizable. It is a test that is blind to the
strategically pivotal difference between reciprocating and unconditional
cooperating. A strategic deontology acknowledges that we cannot uni-
versalize ignoring the fact that the exercise’s point is to identify maxims ﬁt
for members of a kingdom of players – beings who decide for themselves.
Therefore, in a strategic world, imagining yourself unilaterally making
the choice between everyone cooperating and everyone defecting is noth-
ing like imagining yourself choosing for everyone in situations relevantly
like yours. The essence of your situation is that you are not choosing for
everyone.
So, my proposal is: treat strategic deontology as an alternative to ‘act-
deontology’ and envision a choice among strategies, not actions. Maxims
like ‘I should cooperate’ versus ‘I should free-ride’ miss the moral core of
your alternatives. Instead, describe your alternative maxims as ‘I should
encourage partners to cooperate’ versus ‘I should encourage partners to
free-ride.’ Now you see that what is properly universalizable is acting so as
to teach your partners to grasp their place in a kingdom of ends and
thereby mature in the direction of moral worth. Teach them to
cooperate.2
(b) We might observe, similarly, that Peter Singer’s interpretation of
the principle of utility – that we should sacriﬁce to a point of marginal
disutility – is not straightforward. What I call parametric utilitarianism
rests on an empirical premise: picking the act with the highest utility is like

2
I do not suppose this move solves all of deontology’s puzzles. It does, however, address
some ‘indeterminacy of description’ problems in articulating a maxim’s proper form as the
subject of the universalizability test.
134 Chapter 8

picking the outcome with the highest utility.3 Given that supposition, the
only thing to consider is which of our two options, give versus don’t give,
has more utility. If giving has more, then give. Keep giving until not giving
would have more utility.
But the essence of a strategic world is that it does not give us that
supposition. It is not an a priori truth that the action with the highest
number leads to the outcome with the highest number, and in strategic
situations it is a howling non sequitur. The numbers that count are not
numbers attached to available acts, but numbers attached to possible
outcomes, where outcomes are consequences not of particular acts, but
of patterns of cooperation. In a strategic world, someone who cares about
consequences aims to induce a response. If the ideal response is coopera-
tive, then an ideal move is a move apt to induce that cooperative response.
If consequences matter, then being moral in a strategic world is about
inducing cooperative responses, not per se choosing them. That means
being moral involves knowing when to walk away from the act with the
highest number. In strategic situations, if you want the best out-
come for all, don’t worry about other people’s payoffs; worry
about their strategies.
Scottish Enlightenment theorists focused on the nature and source of
the wealth of nations. They cared enough about consequences to study
what has a history of actually working. They observed that prosperous
societies are places where traders build partnerships around principles of
reciprocity. It mattered to them that, in our world, actions have more than
one consequence, more than the intended consequence, and the conse-
quence you don’t see coming will matter. The kind of act-utilitarianism
Peter Singer incarnated circa 1972 is remarkably inattentive to what has
any robust history of good consequences. It is useless not because it is
obsessed with consequences, but because it largely ignores them.4
When moral theory conceives our world in solipsistic terms, practi-
tioners living in a strategic world have to ignore it, because real morality
requires people to make choices apt for the strategic world they actually
3
In the terms of a Prisoner’s Dilemma payoff matrix, the empirical premise I am warning
against is the assumption that choosing a row (‘cooperate’) is the same thing as choosing
a cell in the matrix (‘mutual cooperation’). In a strategic world, it is nothing of the kind.
One unilaterally chooses an act, a row, not as a way of unilaterally choosing an outcome,
a cell, but rather as a way of working towards an outcome.
4
Amartya Sen identiﬁes himself as within the tradition of Adam Smith. Sen earned his
Nobel Prize for his work on twentieth-century famines, showing that not one was caused
by lack of food. Famine is caused by eroding rights, not eroding soil. When local farmers
lose the right to choose what to grow or where to sell it, they lose everything, and that is
when people starve. This is what Scottish Enlightenment theorists studied: consequences
(that is, long-term cause and effect) not imagined best responses to potted thought
experiments.
Realistic Idealism 135

face.5 To summarize an argument only hinted at here, the premise that

moral problems are ﬁrst of all political problems yields a landscape of
moral theory somewhat unlike – relevantly unlike – what we see in our
ethics textbooks today.

2.2 Political Animals Live in a Strategic World

Rawls says his assumption that bargainers choose for a closed society ‘is
a considerable abstraction, justified only because it enables us to focus on
certain main questions free from distracting details’ (1993: 12).
In principle, this could be a legitimate move. As Alan Hamlin observes
in his chapter in this volume, a map of the London Underground sets
aside nearly everything about London, even distances and scale, so as to
distil the one kind of information that the map’s users seek to glean from
it, namely the sequences of stops making up the network’s lines.
Yet it is easy to slide from ignoring for clarity’s sake to ignoring with
prejudice: setting details aside not because they don’t affect the answer, but
precisely because they do (see also Hope 2010: 135). Although we must
set aside distracting details and focus on the problem, one thing we must
never set aside as a detail is the problem. Suppose an asteroid were about to
collide with Earth. What would be an ideal response? Hypothesis: we first
ask what would be right under ideal conditions. Leading our list of ideal
conditions: ideally, there is no asteroid about to collide with Earth.
Having noted that ideally there is no asteroid, we respond in one of two
ways. Either (1) we strive to make it true that there is no asteroid, or (2) we
do what would be ideal if there were no asteroid. The latter overlooks
what should be an obvious difference between doing what is ideal as
opposed to doing what would be ideal under counterfactual conditions.
I say this should be an obvious point. Clearly it is no such thing, for
overlooking the difference is a repeatedly observed blunder. Tucson’s city
government once sought to manage traffic flows by designating inner
lanes of major roads as one-way lanes toward the city centre during the
morning rush. During the evening rush, the same lanes reversed and
became one-way lanes from the city centre. At off-peak times, inner
lanes reverted to being left-turn lanes. In a world of ideal drivers, it
might have solved the problem. In Tucson, with its daily influx of elderly
drivers not necessarily quick to adapt to novel conventions, where one
indecisive driver is enough to create a dangerous mess, the system was
a recipe for traffic jams, accidents, and road rage.

5
Singer himself is increasingly aware of such strategic considerations. See also Schmidtz
(2015).
136 Chapter 8

In effect, traffic managers set aside the problem. Or, instead of tackling
a real problem, traffic managers solved an idealized problem. Their job
was to optimize traffic flow, but they chose instead to do what would
optimize traffic flow if drivers were ideal. Beware of idealized pro-
blems. In general, doing what would be ideal – if only the problem
were the ideal problem! – is not a way of being a serious idealist.
Here is the kind of idealism that realistic idealists scorn:
‘My solution is a hammer; therefore, the problem ideally would be
a nail. Reality is not a nail, obviously, so no one is saying my hammer
is a real solution. Still, I just proved that the ideal problem is a nail!
Therefore, impractical though my hammer may be in the real world, it
remains an ideal.’
By contrast, to a realistic idealist, saying a traffic management system
would work for ideal drivers says nothing in its favour even as an ideal.
An ideal traffic manager works with an accurate picture of the real
problem.
Some idealizations approximate reality, so some ideal solutions are
approximate solutions to real problems. Tucson traffic managers blun-
dered into misconceiving their ideal solution as an approximate solution
to a real problem. That I am not alone but instead live in a strategic world
of separate agents who decide for themselves is not a distracting detail.
If a proposal stipulates that people will not react to our intervention in the
way human beings do react, then the intervention is not an approximate
solution, or even a real response. If Rawls is right to say ‘an important
feature of a conception of justice is that it should generate its own support’
(1999b: 119), then a serious investigator does not set aside whether
a conception of justice actually has that feature. A serious investigator
checks.

2.3 Solipsism as a Snapshot of Justice

There is a literature on whether Rawls was warranted in assuming ideal
bargainers would fully comply with principles of justice. But consider how
much greater a stretch it is to assume ideal bargainers not only take their
own compliance but the compliance of others as given. Once we cross that
line, we are no longer checking to see whether a conception has the key
feature of being able to generate its own support. Instead, we are imagin-
ing what it would be like not to need to check – not to have a political
problem. In different words, once we cross that line, we are no longer
stipulating simply that ideal bargainers are honest; we also are stipulating
that they are clueless about the human condition’s core feature: that what
Realistic Idealism 137

people need from each other more than anything is to create conditions
under which they can afford to trust each other.
G. A. Cohen’s objection to Rawls is that
[I]f we assume, following Rawls, that individuals are motivated to comply with
justice, then the need to trade off equality and well-being disappears. It only
arises in the ﬁrst place because talented people demand incentive payments to
become more productive. But people who are motivated to realize justice fully
would not demand incentive payments but rather increase productivity with-
out them. (Hamlin and Stemplowska 2012: 57, paraphrasing Cohen 2008)

Perhaps Cohen thought that people motivated by justice would not

demand incentive payments. But even if that were true, the fact remains
that even unshakably motivated Rawlsian bargainers would demand
motivating incentives for the people they represent. Rawlsian contractors
have a tough assignment: they are contracting on behalf of people other
than themselves. Rawls can stipulate that bargainers are whatever bargai-
ners need to be to get Rawls’s desired solution, but bargainers can’t
stipulate the character of human psychology. By assumption, Rawlsian
bargainers know human psychology. Therefore, their moral motivation
does not blind them to the reality of what motivates the classes of people
they represent. They know that the psychology of citizens at large is
exactly what it is. (See also James 2012: chap. 4.)
It is a mistake to think we are imagining what is ideal when we imagine
what would be ideal if compliance were something we got for free, rather than
being the precarious achievement that it is. We are supposed to be theorizing
about how to form a community, hold it together and make it worth holding
together. (Let’s not confuse this with talking about policy as opposed to
theory. To say political theory is theory about what holds communities
together and makes them worth holding together is not to propose
a policy; it is to identify political theory’s subject matter.) Setting aside
compliance problems goes astray not because it bears on ideals, but because
it fails to bear on problems. To say ‘ideally we would not have compliance
problems’ is like saying ‘ideally we would not need to drive defensively.’ It is
a remark about a world whose problems are not like ours. We have a history
of solving compliance problems, but there is no recognizable rendering of
the human condition on which we do not have compliance problems.
To set aside that we live amongst agents – beings who decide for
themselves whether to comply – is to set aside the deﬁning problem of
political theory. If an institution is ideal in a given setting, it is by virtue of
what it leads people to do in that setting. Keep this in mind: what isn’t
an ideal incentive structure isn’t an ideal institution. Whenever we
138 Chapter 8

choose an incentive structure, we choose the compliance problem that

goes with it (Schmidtz 2011a).

2.4 Sometimes, Ignoring Feasibility also Ignores Desirability

When we ask whether we are looking at an ideal campground, we can
ignore ravines standing between us and that supremely desirable camp-
ground. I agree with Estlund (2008: 269) and Cohen (2009: 10) that
ravines bear on whether getting there is feasible, but not on whether
getting there would be desirable. Further and crucially, the cost of getting
there can affect whether striving to get there is desirable, but not whether
being there is desirable.
But here is the key. In the imagined case, I agree that to ignore ravines is
to ignore questions of feasibility, and that we can ignore feasibility and still
be discussing an ideal. Yet we abandon anything recognizably ideal if we
ignore whether a campground is suitable as a place to camp. To ignore what
will befall us if we get there is to ignore not whether getting there is feasible
but whether being there is desirable.
In a Carens Market, to use Estlund’s (and Cohen’s) example, we
imagine everyone being taxed in such a way that everyone ends up with
equal disposable income after taxes. Despite this, by hypothesis, we also
imagine everyone working hard to maximize gross income. Estlund
uses the example to stress: ‘So the fact, if it is one, that we shouldn’t
institute the Carens Market because people won’t comply with it, doesn’t
refute the theory’ that people should comply (2011: 217). Estlund adds, ‘it
is doubtful that the content of social justice is sensitive in this way to
untoward motivational features of people’ (2011: 227).
But we choose how to conceive of justice, and whether we see human
motivation as ‘untoward’ turns on whether the thing we want to call
justice characteristically induces untoward behaviour. If we see that
what we want to call justice has that characteristic feature, that is reason
to stop calling it justice, or at least to stop calling it ideal. If it predictably
would realize our worst potentials as human beings, the relevant lesson is
not that the Carens Market is altogether infeasible, but that as an aspira-
tion it is altogether unworthy.6 It does not solve a problem; it solves an

6
We can say, the true ideal here is not bare instituting, but rather a conjunction of instituting
and complying. So, the actual Carens ideal is a conjunction of ‘make sure work doesn’t
pay’ and ‘workers keep acting as if it does’. To Estlund, the fact that we should not
implement the ﬁrst conjunct when the second is false has no bearing on whether the
conjunction as a whole is ideal – even if, in our strategic world, instituting the ﬁrst conjunct
is a paradigm of what renders the second one false. All sides seem to agree on this much: (1)
The Carens incentive structure by itself is not ideal. (2) At best, it would be ideal only if we
Realistic Idealism 139

idealized problem. The lesson is not that we have no way to get there, but
that we have no reason to want to.
Some idealists think ignoring compliance problems is ignoring some-
thing analogous to whether a campground is reachable. Not so. Ignoring
what an incentive structure would drive people to do is like ignoring
whether a campground would be terrible.

2.5 When Institutions Turn People into Monsters, Blame

the Institutions
Estlund is correct when he says ‘that a standard won’t be met might count
against people’s behavior rather than against the standard’ (2011: 209).
For example, we may predict that students will fail our exam, without
blaming our exam. That point of agreement notwithstanding, the fact
remains that responsible reﬂection on a predictably bad outcome begins
with the role our standards play in bringing it about. That students
predictably misread double negations is not a defect in our exam, but
littering our exam with double negations is.
Of his utopian theory and its postulation of unrealistic standards,
Estlund says:
People could be good, they just aren’t. Their failures are avoidable and blame-
worthy, but they are also entirely to be expected as a matter of fact. So far, there is
no discernible defect in the theory, I believe. For all we have said, the standards to
which it holds people might be sound and true. The fact that people will not live
up to them even though they could is a defect of the people, not of the theory.
(2008: 264)7
Be that as it may, if we give people a system that trips them up, and don’t
want to trip them up, then we don’t celebrate our ability to trip them up by
saying, ‘People could be good at avoiding the trap I set for them; they just
aren’t. That my system turns a normal human trait into a fatal ﬂaw is
entirely to be expected, but that is a defect of the people, not the system.’
Note: saying there is no discernible defect is not the same as saying there
is no defect. If we want to discern whether our ideal is worth a try, then we
will not treat our tools for discerning defects as distracting details. Yet, when

could assume workers will comply. But (3) if we can safely assume anything about worker
compliance, it is the opposite.
What else needs to be said? Perhaps this: anything we have reason to regard as ideal
surely has at least some potential not to be catastrophically misleading as a basis for
practical proposals. There is no such potential in alleged ideals like ‘when work stops paying,
workers keep acting as if it does’.
7
By ‘discernible defect’ Estlund has something like blatant self-contradiction in mind.
Being unﬁt for people as they are evidently is not a discernible defect, but is instead
a defect of the people.
140 Chapter 8

we set aside whether our vision has a robust history of being a hideous
response to the human condition, we are working to make sure our vision
has no discernible defect, while doing nothing to make sure it has no defect.
I may imagine how ideal it would be to move my pawn to K4, but if I fail
to anticipate my partner’s response, then my so-called imagination is, to
chess players, the paradigm of failed imagination. It takes imagination to
be a realist. The player who anticipates what can go wrong is the one
whose imagination other chess players have reason to admire. Imagining
what would be ideal in a parametric world is no substitute for being able to
imagine what is ideal in a strategic world.
Estlund speaks of ‘motivational features that are themselves moral
defects’. Yet only some institutions elicit predictably defective behaviour,
whereas other institutions are exactly right as responses to characteristi-
cally human motivational features. If sexism were an underlying propen-
sity, switched on or off by institutional settings, then we have a duty to
choose institutions that switch it off. If we choose institutions that switch
sexism on, it is our choice of institutions that is reprehensible, not human
nature. When a theorist conceives of justice as answering to a vision rather
than to people, the defect is not in the people. If our vision is poisonous for
people as they are, the right response is to stop blaming people for being ill
equipped to survive what we want to give them, and to start wanting
something else.
Crucially, it is false that people ‘just aren’t good’. How good people are
is variable, sensitive to how their institutional structure handles their
separateness as decision-makers. When it comes to fostering society as
a cooperative venture, it is misleading to say people are not good. Rather,
people are not as good (not as cooperative, not as benevolent, not as
trusting) when operating within frameworks that make free-riding pay.
Pablo Gilabert says, ‘It is part of the job of political philosophy to keep
ambitious ideals clear and visible, and to criticize a political culture
when it becomes complacent and superficial’ (2015: ms). I agree. Yet
the phrase ‘ivory tower’ designates philosophy that is complacent and
superficial, not ambitious. Insisting on tracking evolving reality is one
way – I suspect there is no other – to keep ambitious ideals clear and
visible.
It was not an ambitious ideal that drove G. A. Cohen’s (2003) retreat into
‘feasibility is philosophically irrelevant’ mode, as communism fell apart
before his eyes. If you are discouraged and hate to admit that your case for
communism’s economic superiority has been tested and found wanting,
then you lean towards a particular kind of idealism. By contrast, if you want
to avoid complacency, don’t judge people according to whether they
fit your vision. Judge your vision according to whether it fits them.
Realistic Idealism 141

2.6 Realism and Conservatism

Judging your vision according to whether it is a competent response to the
human condition is not a way of being conservative. Starting from here is
a way of starting, not a way of staying, so realism is an orientation towards
progress, not a form of conservatism. To a realist, reality is not what needs
justifying so much as what needs improving.
Note that the relevant notion of feasibility here is dynamic; what can’t
be done today may one day be within reach. It’s realistic to anticipate that
the ceiling of possibility will someday look very different from how it looks
today. In 1789, William Wilberforce arguably had no way to muster the
votes to abolish England’s slave trade, yet it manifestly was feasible to
work towards a day when England would have the will to abolish it.
We can be biased in an unrealistically conservative as well as an unrealis-
tically radical direction. We underestimate prospects for change at least as
often as we overestimate them.8

3 We Are Diverse
Theorizing about how political animals should live could start by obser-
ving the extent of disagreement and diversity in human society. One
implication of diversity: diversity is only one of many places to start, and
where we start matters.
Consider how idiosyncratic and incompatible our individual visions of
perfection are, thus how unfit they are to be a blueprint for a community.
Part of the essence of toleration, of mature adulthood and of being fit to
live in a community at all is acknowledging that our personal visions do
not obligate others – not even if we are so gripped by confirmation bias
that we can talk ourselves into believing that our visions cannot reason-
ably be rejected.
The most primordial political fact of all is the fact that I am not alone.
I live among beings who decide for themselves. I may feel that people
cannot reasonably reject my deepest convictions about justice. But
they can, and they know it. This fact makes politics what it is, and justice
what it is.
Honestly taking the fact of diversity into account comes down to
grappling with a question like this: ‘what terms of engagement are
8
Was Wilberforce overconfident in the justness of his cause? I think not, but that may be the
wrong question. As I understand, Wilberforce’s opponents were overconfident in the
justness of their cause, as majorities usually are. They talk themselves into feeling right-
eous when they bully those with minority views. To complicate things, majorities are not
always wrong, and may even be right most of the time. But when they are wrong, and are
holding back progress, they will be the last to know.
142 Chapter 8

appropriate for people who do not even agree on which terms of engage-
ment are appropriate?’ The question is not cute. It is the crux of the
human condition. Rushing to treat our own intuitions about perfect
justice as if our intuitions were rationally compelling would be
a paradigmatic way of failing to rise to the level of seriousness that justice
demands.

3.1 Thinking We Should Be on the Same Page Is a Problem

Theorists sometimes assume they have high standards (even when others
can see that they don’t), and console themselves with the thought that
human nature is too imperfect to live up to their high standards. In truth,
the problem is not that other people cannot live up to ‘high’ standards.
The simple reality is that there typically is no reason why they should.
People have visions of their own. Liberalism is the insight that this is not
a problem.
Some theories make it seem important that we cannot reach consensus
on destinations. It is not. What matters is that under favourable circum-
stances we coordinate on norms of traffic management. We have no
history of being able to agree on who has the superior destination.
We have a robust history of being able to agree on who has the right
of way.
Freedom of religion is an example of the latter; we reached consensus
not on what to believe, but on who gets to decide. You need not decide
whether my choice of religion is a good choice. You need only decide
whether it is my choice. People saw that they could ignore the most
colonial and brutal premises of their own religions and philosophies.
What won the day was not a religion so much as people deciding that
religion didn’t have to come up. There is no good reason not to let
everyone decide for themselves.
What grew in the soil of religious freedom was more general than
religious toleration. What flourished was liberalism: the idea that we
need not presume to involve ourselves in running other people’s lives.
Our greatest triumphs in learning to live together stem not from agreeing
on what is correct, but from agreeing to let people decide for themselves.
Freedom of speech has a similar point: not to get more speech or to
promote anointed versions of ‘diversity’, but to stop presuming to decide
as a society.
When discussion is not needed, that fact constitutes success in specify-
ing terms of engagement. We make progress by defining jurisdictions that
respect people who want and need to share the road, but neither want nor
Realistic Idealism 143

need to share (or even justify) a destination. No one must accept being
relegated to a category of persons whose destination is less important.
Thriving communities minimize our need to justify our destination to
others. Indeed, the utility of a traffic management system largely lies in
people not needing to justify themselves. We need not stop at intersec-
tions to justify our destinations. We stop only because it is someone else’s
turn. Underlying a healthy society is a logic of coordination rather than
unified agency. In a healthy society, people’s movements constitute a flow
of traffic that moves smoothly, by virtue of people reaching consensus not
on what their destinations should be so much as on who has the right of
way. No one needs to agree about that. It is enough that we simply expect
the people around us to adjust their expectations to fit with what they
think others expect of them.
Ideally, we want to be able to co-exist with all of our neighbours, not
only the ideal ones. Realistic idealism aims to identify what, if anything, is
observably enabling people to thrive under actual conditions, not merely
ideal ones. When disagreement is inevitable, our worthy ideal is to make
disagreement non-threatening – to make it safe to disagree. Aim not to
minimize disagreement but to minimize the need for agreement.
The ideal of a mature political animal is not to win debates, but to avoid
needing to win. Realistic idealism does not delude us into thinking other
people should be on the same page as we are, and therefore avoids cursing
us with the appearance of a mandate to bully those who see things
differently.
Is there any alternative to consensus as a political aspiration? Is there
a realistic ideal? Perhaps it would be something like balance of power. When
people do not feel that they can safely abuse those with different views and
values, society makes progress.9
It is (a not quite realistic) ideal that political power be justified to all
citizens. No one expects total victory on this front any more than we
expect a war on poverty to culminate in a poverty rate of zero. Respecting
this ideal in practice involves minimizing how unjustified a regime’s
exercise of power is. One legitimate way to do that is to minimize the
cost of exit (Pennington 2017). That is hardly a total victory, but approx-
imate success marks a society as genuinely liberal. Being 100 per cent
justified is not realistic, but it is entirely realistic that exit be a non-
appalling option for any citizen appalled to be subject to a given regime.

9
Of course, liberal politics does not simply leave things where they were. It manages trafﬁc
(dictating that people get to choose their own religion, for example). It does not treat all
destinations as equally valuable. It does try to make sure no one (apart from dangerous
criminals) is left facing a light that never turns green.
144 Chapter 8

4 Justice Is Not a Peak

John Rawls arguably was the most influential social philosopher of the
twentieth century. His greatest work opens with the thought that ‘justice
is the first virtue of institutions’ (1999a: 3), from which we infer that
a theorist’s main task is to articulate principles of justice. Rawls’s sentence
resonates. It is lyrical, poetic, compelling.
But it is not right. Historically, we make progress when we acknowledge
that justice is not the first virtue. The first virtue of social institutions is
that they enable us to be neighbours. In practice, the first thing we need
from social institutions is a settled framework of mutual expectation that
keeps the peace well enough to foster conditions that enable society to be,
in the most rudimentary and non-theory-laden sense, a cooperative ven-
ture for mutual benefit.10
Institutions with this virtue make it safe for us to show up and become
a community, contributing goods and services in reasonable expectation
of reciprocation. They set up society to become the cooperative venture
that Rawls wants it to be. They lay a foundation for a solidarity that frees
us to think about what is fair – starting from here.
When we settle disputes, we don’t get resolution by deciding that our
vision has a right to be colonial, and that we can condescendingly dismiss
rival visions as unreasonable. Instead, real resolution starts by aiming for
real resolution. To be in the grip of a vision – any vision – is problematic.
What we need is not to envision, but to listen. That is, we need politics.

4.1 Peaks Are Not Real, but Pits Are

We each have our own theories and visions about the nature of justice:
perfect justice. But our respective visions of perfect justice are too perso-
nal and idiosyncratic to be a basis for moral life in a social world. It is
implausible that justice is any of our idiosyncratic peaks. A vision is not
the kind of thing that could ever be good at managing trafﬁc among
diverse people.
Justice in practice arguably has no essence, which may be why we still
lack an uncontroversial articulation of such essence. We would need to
have an ideal in mind if there were a destination such that arriving at that
summit is just, while arriving anywhere else is not. Yet there is another
way of looking at it: justice is not a speciﬁc place (or distribution) we need
to get to, and it is not a property, except insofar as it consists of an absence
of properties that make for injustice. Specifying the essence of justice has
turned out to be like specifying the essence of ‘non-circle’. The closest we

10
This is how I read Williams (2005).
Realistic Idealism 145

come is to say that justice is (essentially) absence of injustice. If we some-

how were to rid ourselves of all the grinding, vicious, overreaching med-
dling and bullying in the world, justice would simply be the opportunity to
thrive in peace that was left over. There is no problem to solve unless
people are in one of those pits. Something needs to be done – we need to
be somewhere other than where we are – when, but only when, our
situation has features that make for injustice.
A noteworthy virtue of this perspective is that defining justice primarily
in terms of ‘Thou shalt nots’ (and thereby making justice revolve around
an absence of properties that make for injustice) treats justice as limiting
what we can do with other people’s lives rather than as dictating what we
can do with our own. Justice so conceived leaves moral agents with room
to live lives of their own, which is how justice has to be in order to be taken
seriously as a practical guide to living well.
Pits of injustice are not theoretical visions, but real horrors. Justice on
the ground is about avoiding the pits of slavery, persecution, and sub-
jugation that lead to famine. Flourishing societies give people room to
avoid the pits, pursuing their own personal peaks in their own way, ideally
at no one else’s expense.11 The peak metaphor misrepresents that crucial
aspect of reality. Oppression and misery are real. Conceptions of justice
representing justice as a peak are theoretical constructs. Resources we
spend wrestling society towards our imaginary peak and away from some-
one else’s are wasted. Insisting that justice is a peak, more specifically our
peak, is not what gets us out of the pits.
To be sure, there is such a thing as climbing. My objective here is not to
debunk climbing, but to reflect on what climbing is. When societies
climb, it is not towards a peak. When we climb towards a more just
society, we climb towards an expanding, not a converging, frontier of
possibility – an open rather than truncated future.

4.2 Conﬂict Management

Theorists treat justice as more foundational than conﬂict-resolving rules
of practice, yet judges and other conﬂict management practitioners need
to do the opposite. When judges ignore theorizing about what would be an

11
Our personal peaks will of course have positive content, and may have to do with, for
example, reciprocity, equality, need, or desert. I wrote on such things before (Schmidtz
2006), but did not then doubt that justice has an essence. Moreover, I do not currently
believe we can altogether dispense with these positive elements of justice. We may
exaggerate how compelling our own personal peaks can be to others, and thus may
exaggerate how central a place they can hold as organizing principles for a diverse polis,
but they might for all that remain in some way meaningful and relevant. Honestly, I do
not know.
146 Chapter 8

ideal response to ideal conditions, they are doing the right thing. Their
role is to resolve conflict.
Judges play a role in enabling communities to climb, but see that
climbing begins from where we are. It is natural but thoughtless to think
a judge’s job is to dream about how to do a reset from day one and rebuild
society from the ground up according to a vision of justice. If you buy
a house in the United States, you do a title search. The point is not to
ascertain whether ideals of distributive justice single you out as having the
most weighty claim, but simply to uncover any active dispute over title, or
any unsettled dispute within the past forty years or so. If nothing turns up,
we treat the deed as valid. No one needs reminding that there are no
primordially clean land titles. Ascertaining that no one has disputed a title
in forty years is not a way of giving up on justice; it is a way of getting on
with the kind of justice that can ground society as a cooperative venture.
To philosophers, forty years seems arbitrary, but it is property’s role that
dictates what works as a foundation for cooperative society. To coin
a phrase, foundation follows function. Judges try to formulate simple
rules, in a spirit of equality before the law, that enable litigants to get on
with their lives, knowing how to avoid or minimize future conflict.
Property rights, including rights of self-ownership, are essentially rights
to say no. The right to say no makes it safe to come to market and
contribute to the community, thereby promoting trade, thereby promot-
ing progress. When people have a right to say no and to withdraw, then
they can afford not to withdraw. They can afford to trust each other. They
can afford to live in close proximity and to produce, trade, and prosper
without fear.
However, a right to say no is not a weapon of mass destruction.
The operating idea is having a right to decline to be involved in
a transaction, not to forbid transactions among others. Consider a case
(see Schmidtz 2011b). In Hinman v. Pacific Air Transport (1936),
a landowner sues an airline for trespass, asserting a right to stop airlines
from flying over his property. The court’s predicament: because a right to
say no grounds a system of property that in turn grounds cooperation
among self-owners, it was imperative not to repudiate the right to say no.
On the other hand, much of property’s point is to facilitate commercial
traffic. Ruling that landowners can veto air transport is a red light that
would gridlock traffic, not facilitate it. In Hinman, a property system had
come to be inadequately specified relative to newly emerging forms of
commercial traffic. The plaintiff’s interpretation of our right to say no
implied a right to gridlock air traffic, so the edges of our right to say no
needed clarifying.
Realistic Idealism 147

In Hinman, Judge Haney ruled that the right to say no does not extend
to the heavens, but only so high as a landowner’s actual use. Navigation
easements subsequently were interpreted as allowing federal govern-
ments to allocate airspace above 500 feet for transportation purposes.
The verdict made the system a better solution to a particular problem
confronting Judge Haney’s court, leaving us with a system of rights that
we could afford.
If Hinman had a right to veto peaceful cooperation, he would have
a right to veto progress. If Hinman makes demands of people to whom he
is of no use, and works to be someone whom society would be better off
without, then Hinman’s neighbours will ignore him as best they can and
seek out contributors: partners less intent on making demands and more
intent on having something to offer.12
Property’s purpose in managing commercial traffic (the purpose at
stake in Hinman) has to condition the contours of what we call justice,
not the other way around. Taking justice seriously involves seeing justice
as something that comes second, not first, because taking justice seriously
involves seeing justice as something a society can afford to take
seriously.13
From a mediator’s perspective, the test of theory is how it works in
practice, and in practice there is no progress without negotiation and
compromise, aiming for what everyone can live with. It is one thing to
win. It is another thing to get a result about which no one feels
triumphant but to which all can adjust without feeling sacrificed on
the altar of a vision they do not (and no one honestly expects them to)
share.
Judges have to play fair with the cards they are dealt. Judges can
theorize about cards they ideally would have been dealt, and such
theorizing is not necessarily irrelevant. But the relevance of such theo-
rizing stems from its implications regarding how best to play their actual
cards. While a philosopher’s job involves reflecting on how the world
ought to be in the grand scheme of things, actual governance is the art of
compromise in a world that is not a blank canvas. The practical rele-
vance of political philosophy depends on how well we take our cue from
effective conflict mediators.

12
See Harrison Frye, ‘A Different Camping Trip: Offers, Demands, and Incentives’ (pre-
sentation at Chapman University, 10 July 2015) for the idea that there is a ﬁne line
between an offer (to bring a particular service to market for a price) and what others
perceive as a demand (when they want the service but resent having to pay).
13
We cannot afford to think of justice in terms that will render it obsolete as a response to
tomorrow’s problems. See Rosenberg (2016).
148 Chapter 8

Some questions have no answers until judges sort out what will help
current and potential litigants in particular circumstances to stay out of
court. After judges settle a dispute, citizens go forward not with personal
visions of justice, but with validated mutual expectations about what to
count as their due. Judges get it right when they settle it – when they
establish mutual expectations that leave everyone with a basis for mov-
ing on.
Effective judges know this. To them, having personal convictions
about fairness is not good enough. Judges aim higher, and thereby
settle disputes in a way that philosophers and their theories almost
never do. Philosophers spend their days convincing themselves that
they have enough evidence for their view to justify ignoring the evi-
dence against. Judges spend their days giving litigants a way to get on
with their lives.
I have never been employed as a mediator, but after playing
football in high school, I coached and served as a referee. Our task
as referees was to interpret and apply the rules. With responsibility
came power. With power came a measure of discretion. Our calls
could determine a game’s outcome. Crucially, it was not our place
to prefer a particular outcome. Favouring a team would have been
corrupt. Neither had we any right to prefer games ending in a tie.
That too would have been incompatible with the unobtrusive impar-
tiality that deﬁnes successful refereeing. We had a duty not to aim
for any outcome, not even an equal one. It was not our place to win.
Our aim was to let the players play, and let their futures be of their
own making.

4.3 Corruption
Benjamin Barber notes Rawls’s lack of realism in a stinging remark:
‘When political terms do occasionally appear, they appear in startlingly
naive and abstract ways, as if Rawls not only believed that a theory of
justice must condition political reality, but that political reality could
be regarded as little more than a precipitate of the theory of justice’
(1989: 310). Robert Paul Wolff’s criticism is equally sharp. He sees in
Rawls ‘no conception of the generation, deployment, limitations, or
problems of political power’.
It would require very considerable political power to enforce the sorts of wage
rates, tax policies, transfer payments, and job regulation called for by the differ-
ence principle. The men and women who apply the principle, make the calcula-
tions, and issue the redistribution orders will be the most powerful persons in the
society, be they econometricians, elected representatives, or philosopher-kings.
Realistic Idealism 149

How are they to acquire this power? How will they protect and enlarge it once they
have it? Whose interests will they serve? (1977: 202)
It is indeed startling to see the work of the twentieth century’s most
inﬂuential political philosopher described as ‘startlingly naive’.14 And yet,
upon reﬂection, it is amazing that there is no contemporary philosophical
literature on the idea that power corrupts.
Imagine concentrated power in the hands of the worst ruler in living
memory. Assume what you know to be true: namely, concentrated poli-
tical power actually does fall into the hands of people like that. This has an
important implication. When formulating theories about what is
politically ideal, ask ‘ideally, how much power would be wielded
by people like that?’ and not ‘ideally, how much power would be
wielded by ideal rulers?’ Which of these questions is a genuine question
about the human condition?
One theoretical bottom line is this. The fact that power corrupts bears
on how much power we have reason to want there to be. When we ask
how much good an ideal ruler could do with absolute power, we obscure
this. We are working on an idealized problem, and gravitating towards
endorsing as much power as it takes to realize our vision of true justice.
Yet, among actual corruptible human beings, we ought to regard the raw
power to ram any vision of true justice down people’s throats as the
paradigm of what true justice forbids.
Ideal theory done well cannot be a question of how much power ideally
would be wielded by ideal rulers. Ideal theory done well has to be
a question about how much power ideally would be wielded by the sort
of human being who actually ends up acquiring power in human societies
as we know them (Schmidtz 2015).

14
It would be naïve indeed to suppose for example that, for the sake of fairness, university
resources should be distributed among departments in whatever manner is to the greatest
advantage of the least advantaged department. However, what Rawls actually says is: the
principle applies only to the basic structure. We could simply stipulate this, or we could
argue that the Principle more broadly applied often would fail self-inspection. For
example, would it be to the greatest advantage of the least advantaged to treat rules of
university budgeting as mere summary rules that answer case by case to the Difference
Principle? (Would we distribute grades so as to be to the greatest advantage of the least
advantaged student?) By the lights of the Difference Principle itself, ignoring empirical
aspects of such questions is precisely what we have no right to do when evaluating
society’s basic structure and when evaluating the proper scope of the Difference
Principle’s application. If this is Rawls’s view, then his view has none of the naïveté that
Barber and Wolff ﬁnd in the Difference Principle. The Difference Principle informs one
and only one practice: namely the practice of judging the fairness of society’s basic
structure.
150 Chapter 8

5 Conclusion
This chapter offers several conjectures about what it takes to make
theorizing about political animals worthwhile. Formulated as practical
advice, my conjectures are:
(a) Theorize about players, not pawns. We are political animals
living in a strategic world. To theorize about which institutions are
realistically ideal for political animals, we need to theorize about
which incentive structures are ideal.
(b) Theorize about ideals, but beware of starting with ideals. From
what I observe, theorizing in actual practice spirals between our
articulating of problems and of solutions. Introspectively, it will
seem true that before we were reasoning about the one, there was
a previous stage of reasoning about the other. (See also Philp 2012.)
Inevitably, it will feel right to ask ‘to theorize about x, don’t you need
some conception of y?’ Perhaps observation inevitably is theory-
laden. (That very thought is so obviously a theory-laden observa-
tion.) However, it is just as true that some theories (including some
ideal theories) are observation laden, and those are the theories we
have reason to take seriously. Those are the theories that began life as
responses to something real.
(c) Avoid solving idealized problems. More generally, theorizing
about what would be ideal if reality were no constraint is a variation
on the idealist theme, but not a realistic one. Realistic idealism works
in the space of educated guesses about how the world works and how
the world could work.
(d) Set aside details and focus on what you see as a problem’s
essence. Acknowledge that you exercise judgment when you set
aside details. Even if you do not beg the question, others will think
you did. When you simplify, beware of the impulse to simplify with
prejudice by setting aside, as a ‘distraction’, what reveals that your
solution is not ideal.
(e) Acknowledge that your reasons for seeing the world as you do
are not compelling. Theorize about a world of people who do not
see it your way, and who are perfectly aware that there is no reason
why they should. Societies thrive not when they minimize disagree-
ment so much as when they minimize the need for agreement.
(f) Question the platitude that justice is the ﬁrst virtue of social
institutions. In practice, what a theorist calls justice will be that
theorist’s personal vision. But in a world of people who see things
differently, the ﬁrst virtue of social institutions is that they curb the
hunger to impose a vision. To do that, institutions need to manage
Realistic Idealism 151

traffic in such a way as to minimize conflict and to resolve conflict

effectively when it does occur.
Social structures that make it easier to resolve and avoid conflict go
a long way towards fostering society as cooperative venture. To the extent
that a society is such a venture, it is responding well to the human
condition. People are learning to trust each other far enough, and to
adjust their expectations far enough, to constitute themselves as
a kingdom of ends.
Over-specialized theorists will rush to get to more familiar ground by
pointing out ways in which a society can be thriving yet not just.
Of course! What suffices to resolve conflict is not guaranteed to be fair.
Nevertheless, a resolution that stops the fighting will tend to do so partly
by virtue of resonating with what seems fair enough at the time. Some
societies have a primary, towering liberal virtue – the virtue of letting
people pursue hopes and dreams of their own, in ways that make them
appreciate each other as neighbours. Those societies are not guaranteed
to be just, yet those are the societies that have a chance to be just.

Acknowledgements
My work on this chapter was supported by a grant from the John
Templeton Foundation. Opinions expressed here are mine and do not
necessarily reﬂect views of the John Templeton Foundation. Adrian
Blau’s advice was illuminating and helpful at every turn. The chapter
also beneﬁted from comments by Sameer Bajaj, Harrison Frye, Michael
Huemer, Ellen Mease, Richard Miller, Jacob Monaghan, Joshua Ramey,
and Danny Shahar.

References
Barber, Benjamin, 1989. The Conquest of Politics. Princeton, NJ: Princeton
University Press.
Cohen, G. A., 2003. ‘Facts and principles’, Philosophy and Public Affairs 31:
211–45.
Cohen, G. A., 2008. Rescuing Justice and Equality. Cambridge, MA: Harvard
University Press.
Cohen, G. A., 2009. Why Not Socialism? Princeton, NJ: Princeton University
Press.
Estlund, David, 2008. Democratic Authority. Princeton, NJ: Princeton University
Press.
Estlund, David, 2011. ‘Human nature and the limits (if any) of political
philosophy’, Philosophy & Public Affairs 39: 207–37.
Estlund, David, 2014. ‘Utopophobia’, Philosophy & Public Affairs 42: 113–34.
152 Chapter 8

Gilabert, Pablo, 2015. ‘Justice and feasibility: a dynamic approach’, in Kevin Vallier
and Michael Weber, eds., Political Utopia: Contemporary Debates. Oxford
University Press.
Hamlin, Alan and Zoﬁa Stemplowska, 2012. ‘Theory, ideal theory, and the theory
of ideals’, Political Studies 10: 48–62.
Hope, Simon, 2010. ‘The circumstances of justice’, Hume Studies 36: 125–48.
James, Aaron, 2012. Fairness in Practice. Oxford University Press.
Mason, Andrew, 2010. ‘Rawlsian theory and the circumstances of politics’,
Political Theory 38: 658–83.
Pennington, Mark, 2017. ‘Robust political economy and the priority of the right to
exit’, Social Philosophy & Policy 34: forthcoming.
Philp, Mark, 2012. ‘Realism without illusions’, Political Theory 40: 629–49.
Rawls, John, 1955. ‘Two concepts of rules’, Philosophical Review 64: 3–32.
Rawls, John, 1993. Political Liberalism. New York: Columbia University Press.
Rawls, John, 1999a. A Theory of Justice. Cambridge, MA: Harvard University
Press.
Rawls, John, 1999b. Law of Peoples. Cambridge, MA: Harvard University Press.
Rosenberg, Alexander. (2016). ‘On the Very Idea of Ideal Theory in Political
Philosophy,’ Social Philosophy & Policy 33: 53–75.
Schmidtz, David, 2011a. ‘Nonideal theory: what it is and what it needs to be’,
Ethics 121: 772–96.
Schmidtz, David, 2011b. ‘Property’, in George Klosko, ed., Oxford Handbook of the
History of Political Philosophy, Oxford University Press, 599–610.
Schmidtz, David, 2015. ‘Corruption’, in Subramanian Rangan, ed., Performance
and Progress: Essays on Capitalism, Business, and Society. Oxford University Press,
49–64.
Schmidtz, David. 2006. Elements of Justice. Cambridge University Press.
Schmidtz, David, 2016. ‘After solipsism’, Oxford Studies in Normative Ethics 6:
145–65.
Singer, Peter, 1972. ‘Famine, afﬂuence, and morality’, Philosophy and Public Affairs
1, 229–43.
Sleat, Matt, 2013. Liberal Realism. Manchester University Press.
Smith, Adam, 1984. Theory of Moral Sentiments. Indianapolis, IN: Liberty Fund.
Williams, Bernard, 2005. In the Beginning Was the Deed. Princeton, NJ: Princeton
University Press.
Wolff, Robert, 1977. Understanding Rawls: A Reconstruction and a Critique.
Princeton, NJ: Princeton University Press.
9 Conceptual Analysis

Johan Olsthoorn

1 Introduction
Concepts are the constituents of thoughts – the basic building blocks of
whatever propositions we form. Conceptual analysis is the philosophical
study of those building blocks. It is one of the more powerful tools
philosophers and political theorists have at their disposal; indeed, for
much of the twentieth century, philosophy in the Anglophone world
was all but equated with doing conceptual analysis (Miller 1983: 35–9;
Wolff 2013: 797–801).
Analytical philosophy values above all rigour in argumentation and
clarity in thinking. The former is supported by logic, the latter by con-
ceptual analysis. ‘[T]he very excellence of analysis’, writes John Stuart
Mill, ‘is that it . . . enables us mentally to separate ideas which have only
casually clung together’ (Mill 1989: 114). Besides elucidating and dis-
ambiguating complex ideas, conceptual analysis helps us to sharpen our
thinking by reﬁning and enriching our vocabulary, structuring our the-
ories and guiding moral judgement. Furthermore, conceptual analysis
diminishes risk of miscommunication by improving our grasp of language
and of the ideas we seek to express with it.
The term ‘analysis’ has more than one sense, and so does ‘concep-
tual analysis’. ‘Analysis’ can mean ‘the resolution or breaking up of
a complex whole into its basic elements or constituent parts’ (Oxford
English Dictionary). Conceptual analysis, so understood, aims to
elucidate complex notions by breaking them up into their simpler
component parts. Legal theorists have sought to clarify the enigmatic
concept of property by reducing it to a bundle of analytically separable
rights to a resource – rights to possess, use, manage, transfer, etc.
(Honoré 1987: 165–79; cf. Penner 1996). Social and political
theorists, similarly, have attempted to translate statements about com-
plex phenomena such as corporate personalities (e.g. the nation or the
state) into more lucid ones about the individuals comprising those
personalities.

153
154 Chapter 9

In a wider sense, ‘analysis’ means ‘a detailed examination or study of

something so as to determine its nature, structure, or essential features’
(OED). Conceptual analysis, so understood, is the systematic study of
concepts. This study can take at least four forms: (1) finding a proper
definition for a given term (what does ‘justice’ mean?); (2) hunting
for theoretically relevant conceptual distinctions (is there a conceptual
difference between ‘justice’ and ‘what the state should do’?); (3) explor-
ing conceptual connections between different concepts (does being gov-
erned by a legitimate authority entail a general duty to obey?); and (4)
studying conceptual change (does Locke’s conception of property differ
from modern accounts?).
This chapter offers concrete advice on how to do conceptual analysis in
political theory. I begin by outlining a straightforward and hopefully
helpful way of thinking about terms, principles, concepts and conceptions
(Section 2). I then discuss various kinds of conceptual analysis, together
with ways of doing them, and offer recommendations for concept forma-
tion (Section 3). Section 4 reiterates three prominent worries about the
very possibility of analysing political and value-laden concepts like ‘jus-
tice’, ‘equality’ and ‘liberty’. These worries, I argue, are exaggerated, but
they do suggest tips to avoid common pitfalls. Next, I explore to what
extent an analysis of political concepts can help us solve normative
problems (Section 5). The final section provides practical recommenda-
tions for interpreting conceptualizations of political notions advanced in
the literature. How should we determine what exactly Rawls meant by
‘justice’ and Locke by ‘property’? While the recommendations offered in
this section apply to any interpretive venture, they may be especially
valuable when studying the history of political philosophy (Section 6).

2 Concepts
This section explains what concepts are and what they are not by distin-
guishing concepts from terms (2.1) and from conceptions, principles and
criteria of application (2.2). It also explains what conceptual connections
are and what conceptual change consists in (2.3).

2.1 Concepts and Terms

To be free to do something, e.g. to drive a car, consists in not being liable
to interference by other people, so the Right says. No, the Left responds,
to be really free to do something also requires having the means or
resources to do it. According to the Right, everyone with a valid driver’s
licence is free to drive a car. According to the Left, access to a car is
Conceptual Analysis 155

needed as well. Lacking resources is a way of lacking freedom; it is not just

an impediment to using your ‘formal’ freedom. The poor are less free than
the rich merely by being poor (Cohen 2011a; Van Parijs 1995: 21–4; cf.
Carter 2011).
This debate in political theory, and perhaps others as well, may strike
you as merely terminological. The Right calls one thing ‘freedom’,
the Left another. Potato, potato; let’s talk substance instead. When are
disputes merely verbal? When we are doing political theory, we want our
contributions to go beyond verbal bickering. How do we ensure that our
arguments are substantive?
First of all, we should distinguish terms and concepts. Terms are
words that express or denote concepts. They are common verbal signs
that allow us to convey ideas to each other. Concepts are the basic ideas
that give sense to a term or expression. (Terms that express or denote no
concepts are meaningless.)1 Terms and concepts should be kept distinct
because no one-on-one relationship obtains between the two
(Oppenheim 1981: 3–4). Synonyms – multiple terms for the same con-
cept – are a familiar feature of languages. Think of ‘liberty’ and ‘freedom’.
(Synonyms often have different linguistic roots, e.g. ‘liberty’ is of Latin
origin while ‘freedom’ is Germanic.) Conversely, the same term may
habitually be used to express more than one concept (homonyms).
Often the context clarifies which sense the speaker has in mind – whether
‘bank’ refers to the sloping margin of a river or to a kind of financial
institution. But sometimes conceptual differences are too small for the
context to settle the matter. Philosophical reflection is then called for.
Isaiah Berlin (2002), for instance, argued that there are two concepts of
liberty. He used two catchy labels to distinguish them: ‘negative’ and
‘positive’ liberty.
Once we understand the difference between terms and concepts – the
former are labels for the latter – we recognize that disagreements can be
merely semantic in two ways. First, we could agree on the nature of the
underlying concept while disagreeing about the appropriate label for it.
According to Hobbes, ‘monarchy’ and ‘tyranny’ make for a distinction
without a difference. Both terms, he contends, express the same concept:
a commonwealth in which sovereignty is held by one individual.
They differ only in tone: ‘tyranny’ denotes a monarchy ‘misliked’
(Leviathan 19.2). Debates about appropriate labels are not pointless:
labels can have great rhetorical effect, as this example shows. Still, settling

1
Not every word is a term. Some words, including connectives like ‘and’, ‘neither’ and
‘either’, are not used to express concepts. They serve a merely syntactical function and
hence lack semantic meaning.
156 Chapter 9

on a label rarely constitutes a substantive contribution to theory.

Rather than criticizing people for using a term other than in
your preferred sense, try to engage their underlying ideas and
arguments.
Second, we could be using the same label to express distinct concepts.
The Right has in mind ‘absence of humanly imposed constraints’ when
talking about liberty; the Left thinks lacking resources compromises free-
dom as well. Are these distinct ideas? Hayek thought so, warning against
the ‘confusion of liberty as power with liberty in its original meaning’, as
it would ‘inevitably lead to the identification of liberty with wealth’ (1960:
17). If Hayek is right, then the quibble between the Left and the Right is
merely terminological: the two camps are arguing about distinct concepts
incidentally both branded ‘freedom’. While substantive disagreement is
possible over the value of each concept (using normative arguments), the
current dispute can be resolved by simply introducing new labels, e.g. by
calling the first idea ‘negative liberty’ and the second ‘liberty as power’.
Multivalent terms (homonyms) are a major source of ambiguity. Their
existence seems a shortcoming of a language – a common system of signs
the goal of which is to enable and facilitate both communication and
reflection. Shouldn’t languages have as many markers as there are con-
cepts? Alas, this is impossible: the number of concepts is infinite. To see
why, more needs to be said about what concepts are. Philosophers and
cognitive scientists have proposed numerous theories about what kind of
thing a concept is and about how concepts relate to other concepts and to
language more generally. I shall refrain from summarizing or evaluating
these debates here (for overviews, see Margolis and Laurence 1999,
2014). What I shall do is offer a way of thinking about concepts that
strikes me as particularly helpful for explaining the role of conceptual
analysis in political theory.
Concepts serve a classificatory purpose: they group together individual
objects, both real and ideational, under a common mental heading.
Cognitive scientists typically describe concepts as mental representations
of categories (classes of objects in the world) (e.g. Murphy 2002: 5).
Concepts, so understood, are indispensable tools for navigating the
world. Possessing a concept helps you understand and respond appro-
priately to any new entity of the corresponding category that you may
encounter. Since we understand individual objects, people and events
through concepts, concepts have been called ‘the glue that holds our
mental world together’ (Murphy 2002: 1). Philosophers tend to see
concepts less as mental representations (‘internal ideas’) than as the
shared, abstract and non-subjective classifications expressed by language.
We can use and understand language by virtue of sharing concepts. As the
Conceptual Analysis 157

constituents of thoughts (‘mental sentences’), concepts are also the build-

ing blocks of whatever propositions we form. ‘Concepts are to complete
thoughts, roughly speaking, as words are to sentences’ (Dummett
2010: 19).
Objects can be classified in infinitely many ways, each classification
corresponding to a concept. How elaborate a system of classifications
(a conceptual scheme) we require is determined functionally.
Communication among certain alternative music fans would be severely
restrained without words to express the different auditory experiences of
‘death metal’, ‘deathcore’ and ‘Gothenburg death metal’. For others, one
overarching category suffices: ‘noise’. The concepts we cast to capture and
classify objects can be ever more fine-grained. Some may find it useful to
distinguish between ‘deathcore’ and ‘grinding deathcore’. Adding the
category ‘bluesy deathcore’ is unhelpful as there is no such thing in the
world today. The notion is conceptually open, however: it is no oxymoron.
Indeed, music may develop in such a way that need for this term arises.
The set of objects falling under a concept is called its extension. Thus, the
extension of the concept ‘death metal’ consists of all the objects in the
world properly called ‘death metal’. Note that mental associations
prompted by that word (such as ‘Slayer’, ‘leather jackets’ and ‘anger’) are
not part of the concept of death metal; these are instead called its
connotations.
As representations of categories, concepts are not the kind of things
that can be true or false (in the jargon, concepts are not truth-apt). Calling
the concept ‘liberty’ true or false makes no sense. Rather, concepts give
sense to statements that can be true or false (‘propositions’). Concepts
can be functionally better or worse, however: they can serve our
practical, theoretical and communicative goals more or less effectively.
Moreover, concepts can be incoherent. The concept ‘justified slavery’, for
instance, is incoherent: as a violation of basic human rights, slavery is
never justifiable. Observe that only composite concepts can be incoher-
ent, as one part has to conflict with another part. It is possible for concepts
to be coherent yet to have no real-world application. Think of the concept
‘unicorn’. Homo economicus – the rational man of classical economics – is
arguably another instance of a concept that cannot be rightly predicated
of anything existing in the world. Concepts without real-world referents
need not be meaningless or without practical import. They can function
as ideals, serve theoretical purposes and even have predictive power.
Concepts can also be vague or indeterminate: when they have no clear
boundaries of application (Section 3.2). ‘Elite’ is an example of a vague
concept: the boundaries of the set of individuals on which ‘elite’ can
truthfully be predicated are intrinsically unclear, even if the dimension
158 Chapter 9

of assessment is settled (i.e. elite skier, social elite). Incoherent concepts

are unacceptable; vague concepts should be treated with care.
Further desiderata of concepts are discussed in Section 3.5.
While concepts themselves have no propositional content and are
hence not truth-apt, claims about how language is ordinarily used –
about what words ‘mean’ – can of course be true or false. While nothing
is amiss with the concept ‘minimalist electronic music’, the assertion that
it captures the ordinary meaning of ‘death metal’ is patently false. While
nothing stops you from forming new concepts, giving a new sense
to existing words is permissible only if the audience understands
and accepts it. The meaning of words is socially determined.
‘A language, as a social institution, is external to any individual speaker.
The meanings of its words are not subjectively associated with them: they
are objectively constituted by the common practice of speaking the lan-
guage’ (Dummett 2010: 83). Disputes are rendered merely verbal the
moment one disputant glosses a term in a sense not accepted by her
opponent: each then uses the same label to refer to different concepts.
Hobbes was notorious for such semantic shenanigans:
Mr Hobs is very dexterous, in Confuting others, by putting a new Sense upon their
words, rehearsed by himself; different from what the same Words signiﬁe with
other men. And therefore, if You shall have occasion to speak of Chalk; he’l tell
You that by Chalk, he means Cheese; and then, if he can prove that what You say of
Chalk, is not true of Cheese; he reckons himself to have gotten a great Victory.
(Wallis 1662: 154)

Let us return to the debate we started with, about the meaning of

‘liberty’. Suppose (linguistic intuitions may differ at this point) that
both the Left and the Right use the term ‘liberty’ correctly, i.e. in accor-
dance with common usage. In other words, neither is guilty of ‘conceptual
stretching’ – applying terms to things beyond their normal usage (Gerring
1999: 360; Sartori 1970: 1033–46). This suggests that the term ‘liberty’ is
equivocal – used to express two distinct concepts. We can check whether
that is really so by examining whether either concept is reducible to
the other. According to G. A. Cohen, ‘lack of money induces lack of
freedom, even if we accept the identiﬁcation of freedom with absence of inter-
ference’ (2011a: 174–5). Poverty, he insists, quoting Berlin, is not like ‘a
kind of disease which prevented me from buying bread . . . as lameness
prevents me from running’. Such inabilities ‘would not naturally be
described as a lack of freedom’ (Berlin 2002: 170; cf. Van Parijs 1995:
23–4). ‘Money structures freedom’ since property arrangements, unlike
physical inabilities, are social relationships of exclusion and constraint
(Cohen 2011a: 175). Property relations are essentially coercive and thus
Conceptual Analysis 159

freedom-depriving. People too poor to buy a ticket are not free to board
a train; anyone doing so would be quickly forced out. Notice that Cohen
makes a conceptual claim – ‘a claim about how certain concepts [here
“lacking money” and “freedom”] are connected with one another’
(2011a: 168) – premised on a claim about the coercive nature of ‘prop-
erty’. His argument involves no normative statements, grounded in
principles. Which brings us to the distinction between concepts and
principles.

2.2 Concepts, Conceptions, Principles and Criteria of Application

We use normative or evaluative concepts like ‘just’, ‘true’ and ‘beautiful’
to express propositions about reasons for action, belief and feeling
(Skorupski 2010, 2012).2 Normative concepts are accompanied by nor-
mative principles – ‘general statements about right and wrong’ (Cohen
2011b: 227). To call a person ‘just’, for instance, is to express a kind of
moral approval of her and to assert that she has met a certain norm or
standard: that of being just. The technical term for that norm is ‘princi-
ple’. Normative concepts often require many different principles, each
appropriate for particular situations, actions or agents. Examples of
principles of justice are: ‘valid covenants should be kept’, ‘from each
according to her ability, to each according to her need’, and ‘social and
economic inequalities are justifiable only if they are to the greatest benefit
of the least-advantaged members of society’. The principles mentioned
are principles of justice, rather than of prudence or etiquette, since the
wrong involved in their violation is ‘injustice’. (On principles more gen-
erally, see List and Valentini 2016: 535–6.)
People may agree on the general meaning of a normative concept while
defending mutually incompatible principles of it. For instance, I might
consider the correct principle of distributive justice to be ‘the social
surplus should be divided according to need’, while you favour instead
the principle ‘the social surplus should be divided according to effort’.
Both of us would then agree that distributive justice expresses the idea of
a fair societal division of resources. But each of us endorses a different
standard to determine whether resources are in fact distributed fairly.
Rawls (1999a: 5) has introduced the terms ‘concept’ and ‘conception’
in political theory to flag this distinction. Concepts, on Rawls’s account,
are broader and less specific than conceptions. Conceptions are
2
Philosophers sometimes distinguish between evaluative terms (used to express approval or
disapproval) and normative ones (used to prescribe or proscribe actions). I shall call
‘normative’ any term used to express propositions about reasons for action, belief or
feeling, thus collapsing the distinction between evaluative and normative.
160 Chapter 9

interpretations of concepts developed by adding specifying principles and

criteria. Let ‘distributive justice’ be the concept of ‘a moral norm
governing allocations of benefits and burdens within a society’. As it
stands, this idea is too general to be applicable: principles are needed to
spell out which distributions are just and which not. The same concept
can be interpreted (fleshed out, operationalized) differently, depending
on which principle of distributive justice we endorse (allocation according
to need, effort, etc.). As Rawls puts it, ‘[t]o develop a concept of justice
into a conception of it is to elaborate these requisite principles and
standards’ (2005: 14 n). Multiple mutually incompatible conceptions
may be possible of the same concept. Widespread disagreement over
what justice demands is therefore compatible with agreement on what
the concept of justice consists in. The latter, Rawls suggests, is found in
the accord amidst all the discord – ‘by the role which these different sets of
principles, these different conceptions, have in common’ (Rawls
1999a: 5).
Rawls’s discussion would have been clearer had he managed to dis-
tinguish explicitly between principles and criteria of application.
Endorsement of the same principle of justice by no means guarantees
agreement about which distributions are just and unjust. For we may well
employ different standards to determine whether an object meets
a principle or falls under a concept, and hence whether the corresponding
term should be applied to it. Criteria of application, as such standards are
called, are non-semantic (not part of the meaning of a word). They
instead function as bridges between concepts and objects in the world
(Martinich 2014: 388). Imagine we both favour the same need-based
principle of distributive justice and agree on the meaning of ‘need’. Moral
disagreement would persist if you would judge enjoying annual paid
vacations a human need, while I would consider them at most desirable.
Insofar as our criteria of what counts as ‘need’ differ, we will apply the
same concept to different objects in the world. Which criteria of applica-
tion are best depends on the context and on the purposes of the speaker
and audience.
Rawls’s use of the distinction is further confusing insofar as it suggests
that conceptions of normative notions like justice and fairness are some
kind of concepts. They are not. Rival conceptions of justice are created
through the addition of at least some normative principles. The concept
vs. conception distinction is profitably used to explain a kind of linguistic
disagreement. Suppose we agree that courtesy is a matter of respect but
that you maintain that respect must be deserved while I think it is owed
more or less automatically. Each of us thinks our own interpretation of the
concept best captures the meaning of ‘courtesy’, the way the term is used
Conceptual Analysis 161

in everyday life (Dworkin 1986: 70–2). Rawls, however, uses the distinc-
tion primarily to explain moral disagreement – for instance, disagreement
over what is just. But when I claim that justice forbids inequalities that
disadvantage the worst off, I don’t just make a linguistic claim about how
the term ‘unjust’ is ordinarily used. I also advance a normative claim about
what I think is morally indefensible. Such normative propositions cannot
be validated by pointing to linguistic practices. They require normative
argument. Normative propositions, and the principles they are based on,
can be true or false.3 If a proposed principle of justice fails to adequately
reﬂect our considered moral judgements, then this evinces that the prin-
ciple is incorrect (see Knight’s chapter on reﬂective equilibrium in this
volume). Principles, unlike concepts, are truth-apt. Conceptions of nor-
mative concepts are therefore also truth-apt. They differ in this respect
from conceptions of non-normative notions like courtesy. The latter
cannot be false qua conceptions (although they can fail to capture the
ordinary meaning of the corresponding term).

2.3 Conceptual Connections and Conceptual Change

Concepts are not self-standing entities: ‘concepts always appear in clus-
ters that are mutually deﬁning, sustaining and, for that matter, constrain-
ing’ (Freeden 2005: 125). Some conceptual clusters exist by virtue of
robust patterns of associations: many people seem disposed to associate
‘banker’ with social irresponsibility, for one reason or another.
Philosophically more important are conceptual connections – links to
other concepts entailed by the concept itself, following logically from
the internal structure of the concept. The concept of a claim-right, for
instance, entails the concept ‘obligation’. My claim-right against you
(with respect to x) logically implies that you have an obligation to me
(with respect to x) (Hohfeld 1923: 36–8; Wenar 2013). Conceptual
entailments allow us to make conceptual claims (claims that purport to
be true by virtue of the meaning of terms).
Concepts, we have seen, can be articulated in competing ways. Rival
conceptions often weave a different web of conceptual connections
around the same concept through the addition of different principles
and criteria of application, each with its own conceptual implications.
To conceptualize ‘liberty’ as ‘non-domination’ is to forge a conceptual

3
Most analytical political theorists labour on the assumption that there is such a thing as
justice, and that the theories we develop articulate this ideal for better or worse (propo-
nents of a non-cognitivist metaethics will disagree). Those political theorists may, in turn,
disagree over whether ‘justice’ is a mind-dependent moral property. On moral (anti-)
realism and moral (non-)cognitivism, see Joyce (2015).
162 Chapter 9

link between ‘liberty’ and ‘arbitrary power’ through the idea of ‘non-
domination’ (Pettit 1997: 51–79; 2012: 26–74). ‘Liberty as the absence
of interference’ entails no such link. On the non-domination conception
of liberty, to know whether an individual is free requires exploring
whether she is subject to an arbitrary or uncontrolled power. This require-
ment is a conceptual one: it follows logically from conceptualizing free-
dom in this way. Proponents of rival conceptions are spared this
requirement – though possibly at the cost of linguistic and theoretical
plausibility (Section 3.3).
The way a concept is linked to other concepts is not static and unalter-
able: connections between concepts can be rearranged and change over
time. Even as entrenched a conceptual relation as that between claim-
rights and obligations is plausibly conceived differently. Building on his
interest theory of rights, Joseph Raz (1984) argues that claim-rights are
interests deemed sufficiently important to justify holding others under
a correlative duty (barring conflicting considerations). On this concep-
tion, claim-rights do not simply have corresponding obligations in others:
rather, they are the justificatory grounds for these obligations.
Conceptual change in political theory consists not in changing concep-
tions, but in alterations to the internal structure of the concept itself. Rival
conceptions of normative concepts gain and lose prominence over time.
The Rawlsian conception of justice as fairness seems more popular today
than the utilitarian conception of justice as efficiency (Section 3.4). Such
changing fates do not necessarily imply conceptual change: they may
merely reflect the shifting popularity of the normative principles accom-
panying different conceptions. Conceptual change has occurred when-
ever the concept itself has changed. Such change usually manifests itself
in a rearrangement of the cluster of concepts in which the concept is
embedded and through which it is understood.
For example, in the early modern period distributive justice was
generally regarded as (1) a virtue of individuals, (2) regulating the
division of common goods (including spoils of war and political jobs)
(3) according to merit. Today, we tend to think of distributive justice
as (1) a virtue of social institutions, (2) regulating, among other
things, societal divisions of resources, (3) to which people have
a right (Fleischacker 2004: 1–16). Early modern philosophers did
not simply defend alternative conceptions of the modern concept of
distributive justice. Rather, they conceptualized distributive justice
itself differently. Most of them followed Hugo Grotius (1583–1645)
in holding that distributive justice dealt with ‘merit’ or ‘worthiness’,
rather than with rights (De Iure Belli, 1.1.7–8). Distributive justice did
not endow individuals with (morally) enforceable titles (2.22.16).
Conceptual Analysis 163

Grotius would thus reject the modern conceptual connection between

distributive justice and rights, favouring instead a linkage with ‘worthi-
ness’. Insofar as the idea of justice is conceived differently, judgements
about what is just and unjust will inevitably vary, even if moral
judgements stay constant.

3 Conceptual Analysis
This section explains how to elucidate complex concepts by identifying
and clarifying their simpler component parts (3.1); how to determine the
extension of a concept (3.2); how to explore and expose conceptual
connections between different concepts (3.3); and how to discover pos-
sibly relevant conceptual distinctions through disambiguation (3.4).
The section concludes with recommendations for concept forma-
tion (3.5).

3.1 Conceptual Analysis as Resolution

The term ‘analysis’ is of Greek origin, meaning ‘breaking up’ or ‘unfas-
tening’. Its antinomy is ‘synthesis’, meaning ‘composition’ or ‘putting
together’. Conceptual analysis, in this sense, is a method for elucidating
complex concepts by breaking them up into their simpler and more
comprehensible constituent parts (Beaney 2015; Jackson 1998, 2001).
Locke described the process of analysis as follows:
Justice is a Word in every Man’s Mouth, but most commonly with a very unde-
termined loose signification: Which will always be so, unless a Man has in his
Mind a distinct comprehension of the component parts, that complex Idea con-
sists of; and if it be decompounded, must be able to resolve it still on, till he at last
comes to the simple Ideas, that make it up. (Essay Concerning Human
Understanding 3.11.9)
Suppose we think of justice as ‘obeying the law’. Our idea of justice will
then be ‘confused and imperfect’, Locke claims, unless we have a clear
idea of what ‘law’ and ‘obedience’ mean. To clarify the meaning of
a complex notion, explain the meaning of each term in its defini-
tion. Avoid circularity: Locke’s reductive analysis of ‘justice’ would not
be very informative were he to define laws as ‘just precepts’ (Martinich
2005: 103–4).
Elucidating composite notions by explaining the meaning of their con-
stituent concepts, and their conjunction, is an excellent way of providing
conceptual clarification. A good example in political theory is Carter’s
(2004) analysis of the composite concept of ‘freedom of choice’. He first
164 Chapter 9

deﬁnes ‘choice’ and ‘freedom’ and then constructs a deﬁnition of ‘free-

dom of choice’ out of the two former deﬁnitions. Note that the analysis of
a composite notion requires explaining not only the meaning of its con-
stituent parts, but also of the relationship in which they stand.

3.2 Conceptual Analysis as Extensional Analysis

In a wider sense, ‘analysis’ signifies any ‘detailed examination or study of
something so as to determine its nature, structure, or essential features’.
The classical type of conceptual analysis, so understood, searches for an
encompassing definition of a term by trying to identify the set of condi-
tions that are individually necessary and jointly sufficient for the concept
to apply (the concept’s ‘defining conditions’). Call this approach
‘Extensional Analysis’. Extensional Analysis aims to arrive at
a definition that accurately picks out all instances of a concept and only
those. It thus aims to settle the meaning of a term by determining its
extension (what it properly applies to). This way of doing conceptual
analysis used to be extremely popular. It requires little else but imagina-
tion, some intelligence and a fine armchair – although to be persuasive,
one also needs linguistic intuitions properly attuned to those of the wider
language community.
Extensional Analysis starts by proposing a tentative definition of the
concept you are studying. For this method, a definition ideally takes the
following form: first specify the general class to which the thing to
be defined belongs and then state the specific differences that
mark it off from other species of that same general kind. (This is
known as a differentiating, or per genus et differentiam, definition.) For
example, distributive justice is ‘a moral norm [genus] that governs divi-
sions of benefits and burdens [species]’. Distributive justice as social
justice, then, could be defined as ‘a moral norm [genus] that governs the
division of benefits and burdens [species] within a society [subspecies]’.
To test whether your initial definition accurately captures the
individually necessary and jointly sufficient conditions for the
concept to apply, try to come up with counterexamples.
The counterexamples should be cases that are both disqualified as
instances of the concept by the definition at hand and intuitively plausible
instantiations of the concept. If you deem the counterexample suc-
cessful, tweak and reformulate the definition to accommodate it.
Continue this process until you have identified the defining con-
ditions of the concept. As in the empirical sciences, the definition
arrived at is only provisionally correct: it remains liable to falsification
by future counterexamples.
Conceptual Analysis 165

Suppose we deﬁne liberty as ‘absence of interference’. Is this deﬁnition

of ‘liberty’ adequate? Well, what does it mean to ask this question? On the
current approach, we are asking whether the definition accurately
picks out all and only those instances we deem to fall under the concept
(its extension). Does the proposed definition of liberty succeed in doing
so? Republicans think not, persuaded by the following counterexample.
Suppose we are ruled by an absolute monarch who has the power to
incarcerate us with a twist of her finger. Yet as it happens, her fingers
keep still: we are left alone to do as we like. On the proposed definition of
liberty we are ‘free’ as long as the monarch does not actually interfere with
us. This cannot be right, republicans claim. The proposed definition of
liberty they consider too loose: it identifies individuals as ‘free’ whom
many of us are disposed to believe are not really so. Absence of inter-
ference might be a necessary condition for liberty; it is not, the counter-
example suggests, a sufficient one.
A familiar objection to this style of conceptual analysis is that it
wrongly assumes that concepts have an essence: a common conceptual
core. Wittgenstein has argued convincingly that we have no reason to
expect all instantiations of a concept to share a unique set of common
features that definitions can mark out. Think of all the proceedings we
call ‘games’. Is there any feature common to board games, Olympic
games, drinking games and war games? What binds different instances
of what we call ‘games’ together is perhaps better characterized as
‘family resemblances’: ‘a complicated network of similarities overlap-
ping and criss-crossing’ (Philosophical Investigations §66–7).
No definition or short paraphrase will make sense for all instantiations
of such ‘family resemblance’ concepts. The best we can do is describe the
sundry uses of such concepts.
The idea of ‘cluster concepts’ develops this idea further (Connolly
1983: 14f.). Pointing out that all attempts to specify the individually
necessary and jointly sufficient conditions for the concept ‘nation’ have
hitherto failed, Yael Tamir argues that ‘nation’ is best regarded as
a cluster concept. ‘[I]n order to count as a nation a group has to have
a “sufficient number” of certain characteristics’, such as common history,
destiny, language, territory, religion and ethnicity (Tamir 1993: 65).
Groups may thus qualify as nations by virtue of meeting different criteria.
A second objection Wittgenstein advanced against Extensional
Analysis is that it has trouble dealing with irredeemably vague concepts.
The boundaries of application of vague concepts are indeterminate, such
that they have borderlines cases. Consider the concept ‘tall’. A woman
who is 1.90 meters in height is certainly ‘tall’, whereas a 1.60-meter
woman is not (at least not in our day and age). But can ‘tall’ truthfully
166 Chapter 9

be predicated on a woman 1.75 meters in height? We can stipulate

a criterion – any person at least x centimetres in height shall be called
‘tall’ – but any such criterion will strike us as somewhat arbitrary, precisely
because the concept ‘tallness’ itself has ‘blurred edges’ (Philosophical
Investigations §71). It has been argued that all language is vague; every
proposition would have a certain degree of vagueness (Russell 1923).
A shortcoming common to both this objection and to Extensional
Analysis generally is that they overlook the fact that ‘[t]he criterion for
the correct application of a word is not the same as the meaning of the
word’ (Martinich 2014: 385). Criteria of application are context-
dependent and non-semantic: they can change while the meaning of
a term stays the same (Section 2.2). Thus, while the meaning of ‘tall’
has stayed constant, general improvements in health mean that the
threshold for counting as ‘tall’ has shifted upwards since 1800.
Likewise, the criteria of application for ‘cruel and unusual punishment’
have arguably altered since the US Bill of Rights was ratified, without
concomitant conceptual change. Definitions aim to capture the meaning
of a term as precisely as possible. But whether it is acceptable to use
a word to refer to an object in the world also depends on non-semantic
criteria of application, shared by conversational partners (Martinich
2014: 388).4 Which criteria are reasonable depends on the situation:
whether some activity qualifies as ‘cruel punishment’ will depend on
what we think the point of the prohibition is. The take-home lesson is to
provide clear, precise and plausible definitions as well as opera-
tional criteria of application. In political theory, the latter will usually
depend on wider theoretical considerations (Section 4.3).
These criticisms suggest that proponents of classical Extensional
Analysis have set themselves too ambitious a research programme
(Schroeter 2004). The same is suggested by their conspicuous failure to
produce uncontroversial definitions. Imagining counterexamples may
certainly help us improve our definitions. But the definitions arrived at
do not suffice to determine, in turn, which objects in the world count as
instantiations of a concept. That requires both definitions and non-
semantic criteria of application. However, even if the definitions it arrives
at can never succeed in independently capturing each and every plausible
instance of a concept, Extensional Analysis remains a highly useful
method to sharpen our definitions and to elucidate the ordinary meaning

4
Words denote (mean something), speakers refer to things (Donnellan 1966). A speaker may
legitimately use any word to refer to any object provided the audience understands and
accepts it, without changing the semantic meaning of the word.
Conceptual Analysis 167

of terms. Indeed, as Section 6 argues, even in interpretive ventures

a method resembling Extensional Analysis comes in handy.

3.3 Conceptual Connections

Political theories are systems of political concepts and conceptions (Gaus
2000: 43). Progress in political theory consists in part in the rational
rearrangement of political concepts like justice, liberty and equality.
How we arrange political concepts is partly informed by our moral and
theoretical convictions. Political theories will take different shapes
depending on whether we give normative priority to social equality or to
individual liberty. Other arrangements are conceptually required: if we
conceptualize ‘liberty’ as ‘non-domination’, then our theory of liberty
must explain the concept of ‘being subject to an arbitrary or uncontrolled
power’ on pain of incompleteness. Conceptual analysis aids arrange-
ments of political concepts, and thus construction of political theories,
by exploring conceptual connections between notions.
Philosophical arguments often involve conceptual claims – claims that
purport to be analytically true, i.e. true by virtue of the meaning of terms.
Classical natural law theorists regard it as analytically true that we have
a moral duty to obey any valid law we are subject to. For they deﬁne law as
‘a just precept capable of morally binding subjects’. According to
Francisco Suárez (1548–1617), whenever legislators ‘prescribe that
which is unjust, such a precept is not law, inasmuch as it lacks the force
or validity necessary to impose a binding obligation’ (De Legibus 1.9.4).
Legal positivists reject such strict conceptual connections between law,
justice and moral obligation. In the words of John Austin (1998: 184):
‘The existence of law is one thing; its merit or demerit another. Whether it
be or be not is one enquiry; whether it be or be not conformable to an
assumed standard, is a different enquiry’ (cf. Gardner 2001: 199). For
natural law theorists, to ask whether we should obey a valid law is to
misunderstand the meaning of ‘law’. For legal positivists, this question is
conceptually open: establishing that a norm is legally valid in no way
settles the question of whether we are morally obliged to obey it.
Keeping these two questions distinct is a virtue of legal positivism –
although if natural law theorists are right, it does so on pain of miscon-
struing the concept of law.
Competing theories set out different conceptual arrangements. What
qualiﬁes as a conceptual claim on one theory may therefore not do so on
another. Remember that whether a proposition expresses
a conceptual claim is determined as much by our linguistic prac-
tices as by our theoretical convictions. This holds true in particular
168 Chapter 9

for political theory since it mainly studies theoretical concepts

(Section 4.3). Conceptual claims in political theory are best
regarded as conditionals. If we define slavery as the condition of having
no civil rights, then it follows that citizenship and slavery are incompatible
and that wage slavery is an oxymoron. Were we to define slavery differ-
ently, then this might not follow (cf. Pateman 1988: 146–52).
Notice that Suárez requires no further normative argument to show that
we ought to obey valid law: on his account, this is true by definition.
Whether a proposition expresses a normative or a conceptual claim
greatly affects its truth-conditions. Consider Rawls’s well-known formula
‘justice as fairness’. The expression is ambiguous: Rawls could be read as
voicing either a conceptual or a substantive objection against utilitarianism
(as it happens, he does both). On the first reading, Rawls is saying that
utilitarians like Bentham and Sidgwick failed to grasp the meaning of the
term ‘justice’. They think of justice as ‘a kind of efficiency’ whereas in fact
justice means fairness (Rawls 1999b: 64). On the second interpretation,
Rawls articulates a substantive argument against utilitarianism.
Conceiving of justice as efficiency is wrong because it leads to normatively
undesirable outcomes and/or because it is informed by an unsatisfactory
theory of justice. Thus, Rawls objects that the utilitarian conception of
justice cannot account for our basic moral intuition that ‘slavery is always
[i.e. categorically] unjust’ (1999b: 67). This argument, in turn, strength-
ens his claim that ‘justice’ is best understood as meaning fairness.
Avoid conceptualizing notions in such a way that controver-
sial claims become true by definition. The first reason is pragmatic:
we shouldn’t unnecessarily hamper our thinking about complex ideas
like law and justice by discarding certain views or foreclosing certain
questions by definitional fiat. A second reason is normative: one reason
why we should value justice is because we value fairness and because
fairness requires justice. Were we to define justice as fairness, then we
would trivialize one reason for valuing justice (Carter 2011: 491;
Lovett 2012: 142–4).

3.4 Conceptual Analysis as Disambiguation

Conceptual analysis is an indispensable tool for conceptual reﬁnement, in
turn required for clarity in thinking. Many words are multivalent: used to
express multiple analytically distinct concepts. Whether a term is in fact
multivalent depends on our actual language use and implicit linguistic
commitments. Conceptual analysis helps us to achieve conceptual reﬁne-
ment by disambiguating existing terms and recognizing possibly relevant
conceptual distinctions.
Conceptual Analysis 169

Consider whether the conceptual distinctions you are making

are sufficiently fine-grained for your argument. Many political
terms, including ‘liberty’ and ‘justice’, arguably refer to multiple distinct
concepts, albeit ones that share sundry features. Sometimes it makes no
difference to an argument whether such a term is understood one way or
another. Sometimes it does: your argument about the value of liberty may
be valid only on the assumption that we think of liberty as self-realization
and not as absence of interference. Whenever your argument hinges
on a particular sense of a term, you should state this explicitly and
carefully check your argument for equivocation. If you think your
argument about X holds regardless of how we interpret X, then stating
this explicitly is generally good practice.
To clarify in which sense you are using a multivalent word,
simply stipulate a definition: ‘this paper takes “liberty” to mean . . . ’.
Such definitions are called declarative or stipulative. They are used to
indicate which of a variety of recognized senses of a term you have in
mind. As long as the stipulated definition is a common one, you do not
need to justify your choice – although you should make sure that the
chosen definition meets your argumentative needs and ambitions.
Declarative definitions are functionally distinct from differentiating ones
(Section 3.2). Differentiating definitions do need to be argued for, as they
aim to settle the extension of a concept by determining the concept’s
differentiating properties. Various other kinds of definitions exist, each
serving a separate function (for an overview, see Sartori 1984: 29–34).
How can you see whether an existing term is ambiguous? If you
always try to express yourself in as precise a manner as you can, then
you will probably encounter ambiguities in passing. A more proactive
step is to check a good dictionary like the Oxford English Dictionary –
although dictionaries are generally less helpful for technical terms.
Minimize your use of overly general terms. As Tocqueville
observed, ‘[a]n abstract word is like a box with a false bottom: you
put into it the ideas you want and take them out again unobserved’
(Democracy in America 2.1.16). If you insist on employing highly
abstract terms like ‘positivism’, ‘universalism’, ‘neutrality’ and
‘relativism’, consider adding an adjective to specify what you’re
talking about (e.g. moral relativism, epistemic relativism, relativism
about perspectives).
Another practical tip is to think about what the term’s opposite(s)
are. ‘Good’ is used in both a purely evaluative and a moral sense, as we
immediately recognize once we consider its antonyms (‘bad’ and ‘evil’).
Pateman (1988) has shown that many classical social contract theorists
equivocated over the meaning of ‘civil’. Where they first opposed the civil
170 Chapter 9

to the natural condition, ‘after the original pact, the term “civil” shifts and
is used to refer not to the whole of “civil society” but to one of its parts’ –
namely the public realm, from which women were considered excluded
(1988: 11). Through this equivocation, contract theorists like Locke and
Kant could simultaneously grant and deny women a place in civil society
(1988: 55). Pateman’s conceptual analysis reveals how the natural/civil
antinomy seeped into the private/public one, thus expressing and
legitimizing sexist prejudices.
Finally, you could explore the concept’s internal structure – the
links to other concepts that it presupposes. If diverse usages of the
same term entail different conceptual linkages, then the term probably
expresses multiple concepts. For instance, I believe that two irreducibly
distinct concepts of justice are found in modern political philosophy.
What we may call ‘ordinary justice’ expresses a moral norm ordering
agents to respect moral rights (on pain of injustice). The rights it regu-
lates, including human rights, are logically prior to justice. Ordinary
justice does not deal with the distribution of rights (human rights are
not distributed; we have them by virtue of being human). Distributive
justice, by contrast, is a moral norm allocating rights, e.g. to a share of
societal wealth. Individuals have these rights by virtue of distributive
justice – i.e. because a particular distribution is fair. Rights are thus
logically posterior to distributive justice. Ordinary justice and distributive
justice implicate the concept of rights in divergent ways, which strongly
suggests that they are distinct concepts. My point here is not to convince
you that political theorists unwittingly employ two distinct concepts of
justice, but to illustrate a method for disambiguation.

3.5 Conceptual Design

Sometimes we seek conceptual refinement, not by exploring the
meaning of existing words, but by introducing new concepts alto-
gether. Concepts, we have seen, are really only categories used to
make sense of the world. Like fishing nets, these categories can be
ever more fine-grained – and at times we need finer conceptual
meshes to adequately incorporate our arguments and to properly
channel our moral judgements.
Political theorists often find themselves wanting to refer to a particular
set of objects lacking a common label. Suppose you believe that
a prominent argument in the literature about rights of non-economic
refugees actually applies not only to ‘refugees facing persecution’, but
also to ‘refugees facing systematic marginalization’. The distinction
between the two kinds of refugees is not discovered by disambiguating
Conceptual Analysis 171

an existing term (the meaning of ‘refugee’ remains constant), but by

showing that an argument is unclear about the set of individuals it properly
applies to. As a result, the argument may have hitherto unrecognized
implications. The concept ‘refugees facing persecution or systematic
marginalization’ constitutes a newly formed category – and may warrant
an appropriate label.
Social scientists have written extensively on rules for good concept
formation/design (e.g. Outhwaite 2010; Sartori 1970, 1984). Gerring
(1999: 367) offers nine ‘criteria of conceptual goodness’. Ironically, his
list is itself conceptually confused – some criteria apply to the name
chosen, others to the concept formed. The most important rule Gerring
offers is to introduce new concepts only when you need them for
your arguments or theories. Compare Mill’s (1974: 668) advice: ‘we
should possess a name wherever one is needed; wherever there is anything
to be designated by it, which it is of importance to express.’
Concepts cannot be true or false, but they can be incoherent,
arbitrary, ad hoc, obscure and indeterminate. New concepts should
not be internally incoherent or arbitrary. Arbitrary concepts group
together disparate and apparently unconnected properties, as in the con-
cept ‘bluesy deathcore’. The latter concept is also ad hoc: relevant only to
this particular argument and to nothing else – another vice in concept
formation. Newly introduced concepts should be independently
intelligible. It should be possible to explain what the concept refers to
without invoking your theory. Finally, the extension of concepts
should ideally be determinate, such that it can be known to which
objects in the world, real or ideational, it applies. New terms ideally
express analytically distinct concepts.
Suppose you have formed a new concept, discovered by disambiguat-
ing existing terms or suggested by your theory or arguments. This
prompts a choice of labels – terms that clearly and precisely ﬂag concepts
and distinctions. You could either invent a new word (‘neologism’) or
give an additional new meaning to an existing one (‘neovalent’). This
subsection concludes by offering some guidelines for how to choose
appropriate labels, drawing on Sartori (1970, 1984) and Gerring (1999).
Terms you choose are ideally:
1. Clear, short and memorable. ‘Positive’ and ‘negative’ liberty are
short and memorable labels; alas, they are not very clear. Some labels
are neither clear, nor short, nor memorable. Jeremy Bentham had
a particular talent in coming up with them: ‘archetypation’, ‘phraseo-
plerosis’, ‘chrestomathia’ (1838–43: vol. 8, 126, 247). Avoid acro-
nyms. Introducing catchy names for relevant distinctions is as near
a guarantee for citation fame as there is.
172 Chapter 9

2. In line with common parlance. The meaning of technical terms

inevitably departs somewhat from that of their everyday counterparts
and cognates. ‘Liberal’ expresses a different concept for political the-
orists than for laypeople. Still, to convey what kind of thing a newly
postulated neologism or neovalent refers to, abide as far as possible
by linguistic familiarity. Thus, call a new system to improve educa-
tion, well, anything other than ‘chrestomathia’.
3. Original. Nearly every research field appears to have a distinction
signposted by the names ‘internalism’ and ‘externalism’. Some fields
have several. Multiplication of the same set of technical labels is
a common cause of confusion and miscommunication across disci-
plines and subfields.

4 Analysing Political Concepts

It is sometimes claimed that analysing value-laden political concepts like
‘equality’, ‘justice’ and ‘democracy’ is impossible. People disagree deeply
about what is just and unjust, what democratic and what not. How can we
then determine uncontroversially what the concepts of justice and
democracy mean? Given widespread moral disagreement and conﬂicting
political interests, how should we settle the criteria of application of terms
used to express moral (dis)approval? This section discusses three worries
that have been raised with respect to the analysis of political concepts.
They concern the evaluative nature of many political concepts (4.1); their
allegedly ‘essentially contested’ nature (4.2); and their theoretical back-
ground and ideological use (4.3). None of these difﬁculties, I shall argue,
is insurmountable. They do reveal, however, that studying political con-
cepts requires special handling.

4.1 Essentially and Non-essentially Evaluative Concepts

The word ‘value-laden’ is ambiguous (Carter 2015). One group of value-
laden terms expresses essentially evaluative or normative concepts (recall
that I treat ‘evaluative’ and ‘normative’ as interchangeable). ‘Good’,
‘justice’, ‘reasonableness’ and ‘fairness’ are all essentially evaluative con-
cepts: using these concepts necessarily involves evaluation. To apply any of
these concepts to an object is to evaluate it positively: anyone who
sincerely declares ‘this just arrangement is wrong’ does not know what
justice means. Another group comprises terms that are not intrinsically
evaluative but whose normal use does have evaluative implications.
Examples of such non-essentially evaluative concepts are ‘freedom’,
‘democracy’, ‘charity’, ‘law’, ‘public interest’ and ‘equality’: they are
Conceptual Analysis 173

frequently used normatively, but have sufﬁcient descriptive content to not

necessarily involve commendation or disparagement. It is intelligible,
albeit uncommon, to call acts of charity morally wrong or to reject free-
dom categorically. Non-essentially evaluative concepts like ‘freedom’ are
evaluative not because of any feature of the concept, but because of the
normative commitments of the persons using them (Carter 2015: 284).
Thus, the concept ‘equality’ expresses a value for egalitarians, but not for
malign tyrants. (Egalitarians who morally disapprove of equal distribu-
tions are not necessarily conceptually confused, although they do hold
incoherent normative beliefs.) The widespread negative emotional con-
notations of ‘banker’ are likewise due to speakers’ normative commit-
ments; the term can actually be used purely descriptively.5
The distinction between essentially and non-essentially evalua-
tive concepts helps you keep conceptual analysis and normative
theorizing distinct. Oppenheim (1981) advocates rendering political
concepts suitable for empirical inquiry by treating them as descriptive
concepts, i.e. by defining ‘equality’ and ‘freedom’ in a non-evaluative
manner. This requires a degree of reconstruction: separating the descrip-
tive from the evaluative function of political concepts by creating purely
descriptive definitions. This separation is possible only for non-essentially
evaluative concepts. Definitions are purely descriptive (non-moralized,
value-free) whenever they do not contain any evaluative terms (Carter
2015: 284). Before outlining the advantages of non-moralized defini-
tions, I address Dworkin’s objection that non-normative analyses of
political concepts are impossible: studying political concepts would be
ethical all the way down (2011: 166–70).
Dworkin’s rejection of Oppenheim’s methodological project is
informed by the true claim that few political concepts are value-neutral:
nearly all are ordinarily used to express values (Section 4.3). But since
non-essentially evaluative concepts like ‘freedom’ and ‘equality’ express
values not of themselves but because of the normative commitments of
speakers using these concepts, they can be defined in a value-free manner.
To proffer a value-free definition (a definition without evaluative terms)
does not, however, make a concept value-neutral: for liberals, ‘freedom’ is
valuable however we define it. But again, this positive evaluation is not
part of the concept of freedom, but follows from the normative commit-
ments of liberals. In sum, pace Dworkin, non-essentially evaluative

5
Bernard Williams’s (1985: 129–30) distinction between ‘thin’ ethical concepts (which
lack descriptive content) and ‘thick’ ones (which have both evaluative and descriptive
components) captures roughly the same idea more obscurely: ‘thickness’ and ‘thinness’
are matters of degree, whereas ‘essentially’ and ‘non-essentially’ are not (Carter 2015:
287–9; cf. Dworkin 2011: 180–4; Schefﬂer 1987: 417–21).
174 Chapter 9

concepts like ‘equality’ and ‘freedom’ can be deﬁned in a value-free

manner, essentially evaluative concepts like ‘justice’ cannot.
Avoid moralized definitions of political concepts for at least four
reasons. (1) As Oppenheim argues (1981: 1), scientific dialogue is facili-
tated by the creation of a common vocabulary of operational concepts
that theorists can subscribe to regardless of their normative presupposi-
tions and persuasions. (2) Moralized definitions cause confusion insofar
as they portray moral disagreements as purely linguistic ones (disagree-
ments over the meaning of terms). (3) Moralized definitions may render
controversial questions true by definition. By defining ‘law’ as ‘morally
binding precepts’, natural law theorists like Suárez settled the disputed
question ‘should we obey the law?’ by definitional fiat. Keeping the
questions of ‘what is law’ and ‘what is the moral quality of law’ distinct
is methodologically preferable: to conflate them is to blur the distinction
between conceptual analysis and normative theorizing (Section 3.3). (4)
Finally, moralized definitions provide controversial criteria for applica-
tion. Consider the moralized account of equality advanced by von Leyden
(1963: 67): ‘The true opposite of equality is arbitrary, i.e. unjustifiable or
inequitable treatment.’ By defining one controversial term (‘equality’) by
another (‘inequitable’), von Leyden drastically diminishes the possibility
of agreement on what counts as (un)equal. We may agree that a relation is
unequal while disagreeing about whether the inequality is such that it
renders the relation unjustifiable. Moreover, we might deem the relation
unjustifiable for another reason than its inequality (e.g. because it is
exploitative). (We may agree that something is valuable or justified,
while disagreeing about what makes it so.)
What about essentially evaluative concepts like justice and fairness?
It seems conceptually impossible to define such concepts purely descrip-
tively: to use these concepts is to make a claim about how the world
should be. If so, then given widespread moral disagreement, any attempt
to determine the extension of the concept of justice (the set of objects in
the world properly called ‘just’) will be controversial. This suggests that
studying concepts like ‘justice’ and ‘fairness’ will inevitably be
a normative endeavour, requiring normative arguments.
The worry strikes me as sound. But it does not follow that conceptual
analysis is altogether powerless with respect to essentially evaluative con-
cepts. The objection really applies only to Extensional Analysis
(Section 3.2). Other kinds of conceptual analysis, such as disambiguating
equivocal terms and hunting for conceptual connections, are not
impeded by a concept’s essentially evaluative nature. We can elucidate
what it means to call something just (i.e. what kind of claim we are
making) while disagreeing over which things in the world are just
Conceptual Analysis 175

(Miller 1983: 40–1). While people strongly disagree over what the proper
criteria are for institutions to qualify as just (i.e. over the standards of
justice), disagreement over what makes something count as a standard of
justice (rather than of prudence, good taste or efficiency) is unlikely to run
as deep. This is because recognizing a norm as a norm of justice requires
no normative judgement itself. We can recognize that ‘divide the social
surplus according to need’ is a principle of justice by reflecting on what
the principle aims to achieve (a fair distribution of societal resources).
We do not in addition need to morally evaluate whether the principle
succeeds, in our view, in capturing the true criteria for a fair distribution,
such that it successfully identifies all fair distributions and only those.
This discussion suggests that the extension of concepts like ‘justice’ is
ambiguous. It could either consist of everything in the world that meets
the standards of justice. Or it could consist of all these standards, princi-
ples and norms themselves. When analysing essentially evaluative
concepts like justice, do not conflate the set of objects it evaluates
positively with the set of evaluative standards that fall under the
concept.
An analysis of essentially evaluative concepts can take two forms. First,
reflection on linguistic use may reveal that essentially evaluative
terms are systematically used to refer to distinct kinds of objects.
As Aristotle observed long ago, ‘justice and injustice are spoken of in more
than one way, but because the different senses of each are close to one
another, their homonymy passes unnoticed’ (2000: 81–2). According to
Aristotle, ‘justice’ is used to refer to several different ideas: (1) justice of
persons and (2) of actions; (3) justice as the whole of other-regarding
morality vs. (4) justice as a special virtue; (5) distributive justice; (6)
corrective justice; and (7) justice in exchange. Each constitutes
a distinct concept falling under the umbrella notion of justice.
Disambiguating ‘justice’ in this way is illuminating even amidst intract-
able disagreement over which things in the world are just and unjust.
Second, we can explore the basic conceptual structure of justice
by identifying the elements of justice and their conceptual lin-
kages. These elements constitute linguistic criteria for counting as an
instance of justice. An excellent example is Mill’s analysis of the elements
of justice in Utilitarianism (Gaus 2000: 182–5). After summarizing the
various ways in which ‘justice’ is used, Mill identifies a number of ele-
ments common to each of these usages. This requires analysing con-
ceptually related notions like ‘rights’, ‘desert’, ‘duty’ and ‘impartiality’.
Justice, Mill claims, essentially deals with rights (whether moral or legal)
(Mill 1969: 241–2). ‘Impartiality’ is implicated in justice only insofar as it
is required by ‘the more general obligation of giving to every one his right’
176 Chapter 9

(1969: 243). Try to identify those elements that mark it off from
‘neighbouring’ concepts. How does justice differ from related concepts
like ‘morality’ and ‘expediency’? Unlike expediency, Mill contends, jus-
tice and morality govern duties – things that ‘may be exacted from
a person, as one exacts a debt’, whether by penal sanctions or moral
blame (1969: 246). Justice, in turn, differs from morality in dealing
exclusively with duties with correlative claim-rights: ‘Justice implies
something which it is not only right to do, and wrong not to do, but
which some individual person can claim from us as his moral right.’
By contrast, while we have a duty to act benevolently, ‘[n]o one has
a moral right to our generosity or beneficence’ (1969: 247). Observe
that Mill’s analysis is reconstructive: in everyday language, we do not
always distinguish as clearly between justice and morality.
Conceptual clarification often requires a degree of reconstruction
(for a defence, see Oppenheim 1981: 177–202). The merits of recon-
structive analyses nevertheless depend primarily on their linguistic plau-
sibility, not on their conformity to reflectively endorsed normative
judgements.

4.2 Essentially Contested Concepts

W. B. Gallie has introduced the idea of ‘essentially contested concepts’:
‘concepts the proper use of which inevitably involves endless disputes
about their proper uses on the part of their users’ (1956: 169). Many
political concepts have been claimed to be essentially contested, including
‘distributive justice’ (Gallie 1956: 187), ‘power’ (Connolly 1983: 97–9,
213–25; Lukes 1974: 26–7), ‘the rule of law’ (Waldron 2002) and
‘democracy’ (Gallie 1956: 183–7; List 2011). Disagreement over the
proper application of essentially contested concepts does not simply
reflect conflicting normative judgements. Rather, conflict is inevitable
because of features of the concept itself – ‘features which render contests
incapable of being rationally settled’ even absent moral disagreement
(Swanton 1985: 183).
What makes a concept essentially contested? Consider the concept
‘democracy’. As any essentially contested concept, ‘democratic’ is an
evaluative notion that expresses an internally complex, variously describ-
able and peculiar ‘open’ achievement. People disagree not just over which
institutions are democratic, but also over the proper criteria for counting
as ‘democratic’. Some things, we agree, are paradigmatically ‘demo-
cratic’. An example is our collective, unanimous and informed decision,
arrived at by secret ballot and reflecting our true preferences, to go to Eve
& Adam’s for a pint. However, we may reasonably disagree about which
Conceptual Analysis 177

features make that exemplary decision procedure quintessentially demo-

cratic. For there is ‘no one clearly deﬁnable general use of [“democratic”]
which can be set up as the correct or standard use’ (Gallie 1956: 168).
The idea of essentially contested concepts suggests that describing
a political concept by listing uncontroversial instances of it may not sufﬁce
as an explanation. For we might still disagree about what makes these
undoubtedly democratic decision procedures ‘democratic’. Compare the
construction and development of political ideologies: while almost all
political theorists agree that Locke and Mill are canonical liberal
thinkers, which features of their thought are distinctively liberal is dis-
puted (Bell 2014). It is therefore commendable to state explicitly the
criteria that make something count as an instance of that concept
on your account. This advice makes sense even if the idea of essentially
contestedness does not. As Swanton (1985: 827) has argued, ‘[a]ttractive
though the essential contestedness hypothesis is as a solution to the
problem of intractable dispute in political and moral theory, the hypoth-
esis has not yet been adequately defended.’

4.3 Value-Neutrality and Ideology

Conceptualizations of political notions are rarely value-neutral: their use
‘may imply the superiority of any one of a set of contrasting substantive
ethical points of view’ (Carter 2015: 285). Understanding ‘freedom’ as
the ‘absence of humanly imposed constraints’ favours liberals and liber-
tarians; conceptualizing it as ‘self-realization’ is conducive to communi-
tarians like Taylor (1985). Similarly, thinking of rights in a proprietary
sense (as things we own and can possibly dispose of) befits contractarian
theories more than viewing them as basic interests that warrant societal
protection. A low degree of value-neutrality renders conceptualizations
contested: liberals and communitarians each favour the definition of
freedom that best fits their own theoretical commitments and ambitions.
The more value-neutral you can make a definition, the higher the
chance it will be acceptable to political theorists of different nor-
mative persuasions (Carter 2015: 297). Complete value-neutrality,
though, seems impossible to attain in political theory. What does this
mean for the relation between conceptual analysis and normative theory?
Conceptual analysis may reveal that one contested conceptualization
chimes better with our linguistic practices. However compelling the
communitarian arguments that we should value ‘self-realization’ rather
than ‘the absence of humanly imposed constraints’, such normative
arguments cannot prove that the word ‘freedom’ must be taken to mean
‘self-realization’. The meaning of language being socially determined, we
178 Chapter 9

cannot unilaterally give a new sense to existing terms: deﬁnitions must

bear fidelity to linguistic practices (Section 1.1).
Linguistic practices may be unclear, however, or too variegated to settle
the matter. Conceptual analysis then requires reconstruction. Which
reconstruction (conceptualization) of a notion is preferable will depend
partly on its theoretical plausibility. But how plausible a political theory is
will, in turn, rest on its normative claims and predictions. Normative
considerations will thus indirectly inform our decisions to conceptualize
a notion one way or another. Does it make sense to talk about our initial
rights of moral control over ourselves in terms of self-ownership?
The answer depends, in part, on the theoretical implications of
conceiving the moral relation to our body and person in this way, includ-
ing on the success of resulting theories in capturing and guiding moral
judgement (Section 5).
We should thus attend to the wider political theories within which
concepts are articulated. Rival political theories arrange and concep-
tualize political concepts differently. Insofar as particular conceptualiza-
tions are ultimately preferred on normative grounds, studying political
notions involves theory formation. How to best think of the idea of self-
ownership, freedom or citizenship may itself depend on linguistic as well
as on wider normative and theoretical considerations. Conceptual analy-
sis and normative theorizing are therefore often intertwined in practice.
This renders it all the more important to be clear about whether your
arguments are normative or conceptual in nature. State explicitly
whether you are favouring a particular conceptualization for nor-
mative/theoretical reasons or on linguistic grounds.
The recognition that conceptualizations of political concepts are rarely
value-neutral has inspired the thought that how we use and think of
political concepts is itself a contested political act: political theory
would be ‘political’ both in subject matter and in nature. How concepts
of race and gender, democracy and freedom are understood and figure in
public discourse indeed has tremendous socio-political consequences.
Political concepts are commonly used to express and (re)produce rela-
tions of power and subordination. Critical theorists scrutinize political
discourses, including political theories, in search for recurring (associa-
tive) patterns with oppressive or emancipatory effects for different
groups. Such discourse analysis belongs to the study of ideologies, how-
ever, rather than to conceptual analysis proper (on ideological analysis,
see Freeden 1996, and Leader Maynard’s chapter in this volume).
Although conceptualizations of political concepts are rarely value-free
and practically contested, this in no way spells doom for conceptual
analysis. On the contrary, conceptual analysis greatly benefits discourse
Conceptual Analysis 179

analysis (Freeden 2005: 124). For instance, we can use conceptual

analysis to help us recognize ideological interpretations of con-
cepts by separating the logical core of a concept from theoretical
additions. Take the concept of ‘race’. The term is frequently used to
express the idea that historically and geographically related people have
a shared essence that makes them morally, culturally or intellectually
superior or inferior. Hardimon (2003) has argued that the ordinary con-
cept of race, which marks out groups of humans with shared ancestry by
relevant visible physical features, is not itself essentialist. Its ‘logical core’
does not, therefore, ‘provide a rationale for racism, slavery, colonialism,
or genocide’ (2003: 453). Only so-called racialist conceptions of race,
which attribute a shared and evaluable essence to racial groups, do so.

5 Normative Implications
What follows normatively from a successful analysis of concepts?
Conceptual analysis, I shall argue, is no substitute for normative argu-
ment. It does, however, have important indirect implications for norma-
tive theorizing. Concepts frame our theories and guide moral judgement.
They can also have a distorting effect, however – rendering certain lines of
argument salient while side-lining others.
Normative arguments, and moral judgements generally, need concep-
tual guidance. Conceptual analysis helps normative judgements
target the right propositions, such that they don’t spill over into
adjacent conceptual territory (where they don’t hold true). Mill notes
that proponents of laissez-faire often make arguments ‘far outstretching
the special application made of them’ (Principles 5.11.1). Laissez-faire
advocates object not to government intervention as such, but to coercive
government intervention. Their arguments against government meddling
do not challenge ‘unauthoritative’ state interventions, such as the public
provision of goods the government considers too important to entrust
solely to the care of the market. To avoid argumentative spillover,
consider what is at stake normatively.
Another example is found in Cohen’s article ‘How to do political
philosophy’ (2011b: 227). Cohen urges political theorists not to conﬂate
the following three questions:
1. What is justice?
2. What should the state do?
3. Which social states of affairs ought to be brought about?
Even if justice is ‘the ﬁrst virtue of social institutions’ (Rawls 1999: 3), it
does not follow, Cohen argues, that the concept ‘what the state should do’
is equivalent to ‘justice’. Justice is the primary but not the only virtue of
180 Chapter 9

social institutions – we also evaluate institutions in terms of their efﬁ-

ciency, provision of security and material welfare, securement of privacy,
etc. Do not blend concepts together just because they all signify
things we value. ‘Everything is what it is: liberty is liberty, not equality or
fairness or justice or culture, or human happiness or a quiet conscience’
(Berlin 2002: 172). Furthermore, we could coherently deem an egalitar-
ian division of resources the morally optimal state of affairs without
wanting a coercive entity like the state to bring this about (e.g. because
of the way the state would do so or because we refuse it the requisite
powers). Thus, the claim that it is not the business of the state to redis-
tribute resources does not imply that justice requires no redistribution.
Conceptual analysis helps prevent such argumentative sliding, thus chan-
nelling or blunting the force of particular arguments.
Remember that concepts can be coherent but still theoreti-
cally unwarranted. Does it make sense to conceptualize rights to
autonomy and bodily integrity as ‘self-ownership’? On Locke’s
account, ‘property’ refers to exclusive rights both ‘in their Persons as
well as Goods’ (Second Treatise §173). This broad conception of
property is out of step with modern linguistic sensibilities; most peo-
ple, I suspect, would consider ‘self-ownership’ a metaphor.
Contemporary Lockeans may nonetheless have good theoretical reason
to depart from ordinary language at this point. They claim that we
morally control our body and our person in the same way as we
morally control material resources belonging to us. It does not follow
that the concepts ‘self-ownership’ and ‘property rights in things’ are
structurally identical. They are evidently not: property rights in things
are acquired rights, while the right in one’s person is held by virtue of
being human. Whether the broad Lockean conception of ‘property’ is
acceptable depends in part on whether the conceptual differences
between rights in one’s person and rights to material resources are
theoretically signiﬁcant. A theory that treats two different things alike
(by subsuming them under one category) may well fail with respect to
one of them or to both. Differences in the logical structure of the two
rights may imply that what is true of one is false of the other.
Kant rejected the notion of ‘self-ownership’: ‘Man cannot dispose
over himself, because he is not a thing. He is not his own property’
(1997: 157). While we may sell our house, we are not entitled, according
to Kant, to sell ourselves or parts of our body. Why not? Kant replies that
the notion of self-ownership is internally contradictory: ‘it is impossible,
of course, to be at once a thing and a person, a proprietor and a property
at the same time’ (ibid.). Cohen accuses Kant of ‘trying to pull
a normative rabbit out of a conceptual hat’ here (1995: 212). Kant
Conceptual Analysis 181

does not show why the concept of self-ownership is incoherent, i.e. why
it is logically impossible for persons to own themselves. Even if he did
succeed in showing this, it is better to let moral judgements and
arguments rather than conceptual legerdemain do the norma-
tive weightlifting. Thus if we believe, like Kant, that prostitution is
morally unacceptable, then we should put forth normative arguments in
support rather than claiming that it is conceptually impossible for per-
sons to commodify their body.
Do not mistake the conceptual possibility of a notion with the
actual existence or normative desirability of what it denotes.
Proving a concept to be coherent is one thing; proving that it has real-
world application and/or denotes something normatively desirable is
something else. Proponents of the concept of self-ownership should do
more than dislodge the Kantian objection of incoherency. They should
also show that it is morally true that we own ourselves, e.g. by appealing
to the widespread intuition that ‘we – and not others – are morally in
charge of our bodies and persons’, and by showing that other property
rules like trusteeship or guardianship cannot satisfactorily account for
that intuition (Vallentyne, Steiner and Otsuka 2005: 208). As Cohen
puts it: ‘the concept of self-ownership is not identical with the thesis of
self-ownership: the latter might be false, while the former, being
a concept, cannot be false’ (Cohen 1995: 209).
When analysing evaluative concepts, separate the questions
‘what is X’ and ‘what makes X morally valuable/objectionable’.
A conceptual analysis of racism cannot by itself explain what is morally
objectionable about racism: normative judgement is needed as well.
However, the more reﬁned our grasp of this concept, the easier it becomes
to determine which features make racism wrong. Conversely, if our
analysis focuses exclusively on those features of racism we deem morally
objectionable, then, constricted by our moral judgements, we are in
danger of arriving at an impoverished account of racism.
Section 2.2 pointed out that conceptions of normative notions gener-
ally include normative principles. Consequently, disputes over rival
normative conceptions cannot normally be settled by conceptual
arguments alone. Normative arguments are needed as well. Take
Galston’s (1995) distinction between conceptions of liberalism based
on ‘autonomy’ and ‘diversity’. Galston does not argue that one of these
conceptions captures more accurately the ‘true meaning’ of liberalism –
wisely, as the dispute is not solely about linguistic intuitions. Instead he
argues that the diversity conception better captures what we (ought to)
ﬁnd valuable in liberalism.
182 Chapter 9

Do not mistake relabelling a concept for a normative argument.

Since freedom is ethically valuable, it is morally significant whether lack-
ing resources is a way of being unfree. C. L. Stevenson has called
redefinitions that do nothing else but redeploy emotionally charged
words ‘persuasive definitions’. A persuasive definition ‘gives a new con-
ceptual meaning to a familiar word without substantially changing its
emotive meaning, and which is used with the conscious or unconscious
purpose of changing, by this means, the direction of people’s interests’
(Stevenson 1938: 331).6 Hobbes masterfully understood the nature and
use of rhetorical redescriptions (Skinner 1996: 316–26, 338–43). He
argued, for instance, that the amount of freedom citizens have is deter-
mined solely by the stringency of the law, and not at all by the nature of
their political regime (democratic, dictatorial, etc.). Having a voice in
who governs you is not part of the Hobbesian concept of liberty
(Leviathan 21.6–9).
Rhetorical redescriptions risk being merely semantic. It does not follow
that they are ineffectual. Adopting a particular conceptualization
can make certain arguments and considerations salient, while
side-lining others. As Anne Phillips acutely notes:
thinking of oneself as in a relationship of ownership to one’s body does not
commit one, in some inevitable chain of consequences, to supporting the
commodification of the body and promoting markets in bodily services and
parts. What it does do, however, is skew the kinds of arguments that
become available when societies are considering policies on body matters.
Property claims make the individual property owners the centre of attention
and establish their preferences and choices as the predominant concerns.
(Phillips 2013: 142–3)

How we present an idea – including, as Stevenson reminds us, what we

call it – can distort normative theorizing by obscuring relevant considera-
tions and increasing the salience of others.
Critical theorists have been especially attentive to the argumentative
biases prevailing conceptualizations of political notions have given rise to,
biases often favouring the socio-political status quo. One way to criticize
existing conceptualizations is by constructing ‘a genealogy of concepts’ –
unearthing the contingency of and ideological prejudices implicit in mod-
ern conceptualizations by tracing their history. While it goes beyond the
scope of this chapter to discuss this method exhaustively, the ﬁnal section
will outline a key element of it.

6
Stevenson opts for the term ‘emotive meaning’ because of his non-cognitivist metaethics.
His point applies to evaluative notions generally and does not depend on a particular
metaethical position.
Conceptual Analysis 183

6 Analysing Conceptualizations
As political philosophers, we do not just analyse political concepts them-
selves, as if they were free-floating entities (‘what is justice?’). We also try
to understand, if only preliminarily, how people have conceptualized these
notions (‘what did Rawls mean by justice?’). This section provides con-
crete advice on how to best reconstruct conceptualizations advanced in
the literature. My advice, partly modelled on Extensional Analysis
(Section 3.2), may be especially welcome when studying the history of
political philosophy. Today’s theorists usually try to advance conceptua-
lizations that match our linguistic intuitions (indeed, they are well-
advised to do so!), rendering the task of figuring out what they mean
with a term comparatively easy. This task becomes harder the further we
go back in history. The conceptual landscape has changed dramatically
since Locke wrote the Two Treatises. As a result, his notion of ‘property’
may well depart significantly from ours. Reading historical texts is, in
many ways, an exercise in translation.
How can we best reconstruct conceptualizations advanced in political
texts? The first and perhaps most important tip is to read with an open
mind. Do not simply assume that what Locke meant by ‘property’ is the
same as what you think it means. This also holds for contemporary texts:
G. A. Cohen’s conception of ‘justice’ may well be miles apart from what
you and I think the term signifies. Do not read into the text what is not
there. In Leviathan, Hobbes declares that ‘[t]he notions of Right and
Wrong, Justice and Injustice have . . . no place’ in the condition of ‘warre
of every man against every man’, as it exists outside the state (Leviathan
13.13). Hans Morgenthau, doyen of the realist school in international
relations theory, consequently attributes to Hobbes the ‘extreme dictum’
that ‘the state creates morality as well as law and that there is neither
morality nor law outside the state’ (Morgenthau 1952: 34). But Hobbes is
speaking only about justice. Justice and morality (‘natural law’) are not
the same thing for Hobbes; indeed, natural law exists and is to a limited
extent operative outside the state (Olsthoorn 2015: 20). The distinction
between justice and morality may be irrelevant for Morgenthau’s pur-
poses; it is not for Hobbes.
Read the text, as far as possible, in the original language. This
advice should not be taken to extremes: if your Latin is as rusty as mine,
reading modern editions is likely to be vastly more profitable than sweat-
ing over the original. Remember that you are reading a translation
(when doing so). Even critical editions may translate terms in
a misleading or inconsistent manner. Cross-checking the original can
save you much time and confusion. I have spent ages trying to make
184 Chapter 9

sense of Hobbes’s comment that ‘by natural law, one is oneself the judge’
(De Cive 1.9) – only to discover eventually that Tuck and Silverthorne
(1998) had mistranslated the original Latin: ‘by natural right’ (jure natur-
ali). (For Hobbes, ‘right’ and ‘law’ are opposites – e.g. De Cive 14.3;
Leviathan 14.3, 26.44.)
Stick as far as possible to the original terms. Substituting your
own terms is potentially distorting. Translation inevitably leads to loss
or change of meaning. Even if terms have roughly the same extension –
i.e. when they refer to the same objects – their meaning may differ.
What early moderns called ‘natural law’ picks out roughly the same
objects as our ‘morality’, but the conceptual linkages of the two
notions differ considerably. To highlight one salient difference, early
modern natural laws are moral norms applying to us by virtue of the
kind of beings we are. Modern ‘morality’ lacks such an explicit
grounding in human nature. While substituting your own terms
increases the risk of overlooking conceptual changes, sticking to the
original terms is no panacea: serious historical research is needed to
recover lost connotations and conceptual connections of notions like
‘natural law’. If you are using an original term whose meaning
has changed over time, then describe its meaning in present-
day terms as well as you can.
Be attentive to unexpected usages of the term. We are inclined to
think of ‘property’ as a general term for rules governing access to and
control of material resources (Waldron 1988: 31). Undiscerning readers
of the Second Treatise may assume that Locke was primarily concerned
with safeguarding citizens’ existing rights to material resources against the
State. But Locke uses the term ‘property’ in a far more expansive manner
as including ‘Lives, Liberties, and Estates’ (Second Treatise §123). Such
unexpected usages – lives as property? – should be red flags, indicators of
conceptual change.
What is the best way to systematically reconstruct what a particular
philosopher meant by a term? A good first step is to collect explicit
definitions and descriptions. Locke declares that it is of the nature of
property ‘that without a Man’s own consent it cannot be taken from him’
(Second Treatise §193). Notice that this is not, strictly speaking,
a definition, but rather a description of an essential feature of Lockean
property. Elsewhere, Locke writes: ‘the Idea of Property, being a right to
any thing’ (Essay Concerning Human Understanding 4.3.18). We may for-
mulate a provisional definition: ‘property is the exclusive moral control
over a material or immaterial object.’ To further improve the definition,
consider assembling contextual evidence. The Oxford English
Dictionary often lists earliest possible usages of senses of a term still in
Conceptual Analysis 185

vogue. Specialized etymological dictionaries provide more detailed infor-

mation. When bogged down by an odd usage of a term, it is worthwhile to
look at contemporary translations; the translator’s choice is rarely author-
itative but potentially revealing. Reading contemporary writers,
especially critics, is often illuminating as well. Conceptualizations
may be taken over from earlier thinkers. Critics frequently challenge
innovative conceptualizations, clarifying along the way how they depart
from standard accounts in the period.
The next step is to look for other occurrences of the term, and
check whether they match the provisional definition. Indexes are
handy; word searching a digital edition of a text is more convenient still.
Does the way in which the term is used in the context make sense on the
proposed definition? If not, reformulate the definition. Conversely, reflect
whether on the proposed definition certain things would count as ‘prop-
erty’, and then check whether Locke indeed calls them so. The aim is to
arrive at a description of Locke’s conception of property that accurately
captures its defining conditions. Notice that this method follows roughly
the same format as conceptual analysis as Extensional Analysis, outlined
in Section 3.2. A main difference between interpretative and systematic
analysis – that is, between analysing conceptualizations and analysing
concepts – is the criterion of application. When doing conceptual analy-
sis, we use our own linguistic intuitions and theoretical convictions to
determine whether the case at hand is best regarded as an instantiation of
the concept. When analysing conceptualizations, we use textual passages
to determine their meaning and extension.
This leads to a second difference: inconsistent or incurably ambiguous
usages and incoherent conceptualizations are real possibilities in inter-
pretative ventures. Philosophers are, alas, fallible creatures. They may not
always strictly adhere to their own definitions, and their conceptualiza-
tions may fall short of the ideals of clarity and coherency (Skinner 2002:
67–72). Do not assume perfect terminological consistency.
Terminological inconsistencies can have various causes, including
shoddy thinking. Inconsistencies are especially prone to occur when an
author advances a novel conceptualization – historically and philosophi-
cally the more interesting cases. Be aware that a term may be used in
more than one sense. Locke employs ‘property’ both in a narrow sense
(as rights to material resources) and in a broad sense (as anything a person
may exclusively call her own, including life, liberty, chastity and material
goods) (Olivecrona 1975). Every instance of property in its narrow sense
is an instance of its broader sense; the reverse is not true. It is worthwhile
to explore whether the apparent ambiguity can be resolved at
186 Chapter 9

a higher level by searching for a general deﬁnition that encapsulates

both senses of the term.
Check for theoretical consistency. Like a real scientist, try to falsify
your provisional definition by analysing whether it can fulfil the theoretical
role it is supposed to fulfil. Are Locke’s arguments about property logically
valid on the proposed definition? If they are, then this is evidence that your
definition is correct. It does not follow that your definition is incorrect if
they are not: the problem might be Locke’s rather than yours. Tests for
theoretical consistency are also possible across thinkers. Early modern
philosophers generally argued that people are naturally equal because they
are born free. ‘Freedom’ was understood as ‘being your own master’.
If everyone is by nature their own master, then there are no natural inequal-
ities of hierarchy (of rule and subjection). Thus conceptualized, ‘natural
freedom’ implies ‘natural equality’. We can check the soundness of this
interpretation by examining whether any early modern philosopher
rejected natural equality despite holding that humans are born free.
My final piece of advice is to be wary of interpretive charity. Rawls
is a well-known proponent of this principle: ‘to present each writer’s
thought in what I took to be its strongest form’. Not, he clarified, ‘what
to my mind they should have said, but what they did say, supported by
what I viewed as the most reasonable interpretation of their text’ (Rawls
2007: xiii, also 52; cf. Frazer 2010). If we take ‘reasonable’ to mean
avoiding attributing flagrant inconsistencies and non sequiturs to an
author, few would object: if an argument found in ‘the classics’ is logically
invalid on your reading, consider it a sign of misinterpretation. Rawls goes
further, however: we should interpret a thinker’s arguments and doctrines
in the way that makes most sense to us – in its intellectually strongest
form, the most plausible reading the text can bear. If you are mining the
classics for insights to solve today’s philosophical quandaries, then this is
an excellent principle of interpretation. If you are trying to uncover the
historical meaning of a text, then you should interpret conceptua-
lizations in the historically most plausible way, rather than the
way we nowadays would think most reasonable. Otherwise we risk
forgoing from the outset a major benefit of historical research: unearthing
radically different ways of conceptualizing familiar notions that reveal the
contingency and tacit assumptions of their modern counterparts.

7 Conclusion
This chapter has provided a range of advice, from the obvious to the
controversial, on how to do conceptual analysis in political theory. Rather
than reiterating my main recommendations, I will conclude by
Conceptual Analysis 187

highlighting two methodological assumptions underlying my advice.

Throughout this chapter I have argued that we should not conﬂate
two key tools political theorists have at our disposal: conceptual
analysis and normative theorizing. What does the difference between
the two reside in? Conceptual analysis, I have assumed, is essentially
linguistic analysis: it consists in scrutinizing, systematizing and develop-
ing the classiﬁcations implicit in ordinary linguistic practices. Normative
theorizing, by contrast, consists in the construction of arguments and
theories based on normative rather than conceptual considerations.
Conceptual analysis and normative theorizing thus come with different
felicity conditions and burdens of proof.
The second presupposition is that we should conduct conceptual
analysis independently of normative theorizing. This view is con-
troversial with respect to the many political concepts normally used
prescriptively rather than descriptively. I have argued that conceptually
analysing concepts that express or are normally used to express values is
possible, but requires special care. In particular, it requires distinguishing
between the objects in the world to which an evaluative notion applies
(which always involves normative judgement) and the evaluative norms
or standards falling under that concept (whose nature can be analysed
conceptually).

Acknowledgements
Donald Bello Hutt, Adrian Blau, Benjamin De Mesel, Alex Douglas,
Robin Douglass, Cressida Gaukroger, Siba Harb, Michael Jewkes and
especially A. P. Martinich provided very helpful comments on previous
versions of this chapter. I would also like to thank the KU Leuven
research group in political philosophy (RIPPLE) for a stimulating discus-
sion of an earlier draft.

References
Aristotle, 2000. Nicomachean Ethics, ed. Roger Crisp. Cambridge: Cambridge
University Press.
Austin, John, 1998. The Province of Jurisprudence Determined. Indianapolis, IN:
Hackett.
Beaney, Michael, 2015. ‘Analysis’, in Edward Zalta, ed., The Stanford Encyclopedia of
Philosophy, http://plato.stanford.edu/archives/spr2015/entries/analysis. Accessed
1 September 2015.
Bell, Duncan, 2014. ‘What is liberalism?’, Political Theory 42: 682–715.
188 Chapter 9

Bentham, Jeremy, 1838–43. The Works of Jeremy Bentham, 11 vols., ed. J. Bowring.
Edinburgh: W. Tait.
Berlin, Isaiah, 2002. Liberty: Incorporating Four Essays on Liberty. Oxford University
Press.
Carter, Ian, 2004. ‘Choice, freedom, and freedom of choice’, Social Choice and
Welfare 22, 61–81.
Carter, Ian, 2011. ‘The myth of “merely formal freedom”’, The Journal of Political
Philosophy 19, 486–95.
Carter, Ian, 2015. ‘Value-freeness and value-neutrality in the analysis of political
concepts’, Oxford Studies in Political Philosophy 1: 279–306.
Cohen, G. A., 1995. Self-Ownership, Freedom, and Equality. Cambridge University
Press.
Cohen, G. A., 2011a. ‘Freedom and money’, in On the Currency of Egalitarian
Justice, and Other Essays in Political Philosophy. Princeton, NJ: Princeton
University Press, 166–92.
Cohen, G. A., 2011b. ‘How to do political philosophy’, in On the Currency of
Egalitarian Justice, and Other Essays in Political Philosophy. Princeton, NJ:
Princeton University Press, 225–35.
Connolly, William, 1983. The Terms of Political Discourse. 2nd edition. Oxford:
Martin Robinson.
Donnellan, Keith, 1966. ‘Reference and deﬁnite descriptions’, The Philosophical
Review 75, 281–304.
Dummett, Michael, 2010. The Nature and Future of Philosophy. New York:
Columbia University Press.
Dworkin, Ronald, 1986. Law’s Empire. Cambridge, MA: Belknap Press.
Dworkin, Ronald, 2011. Justice for Hedgehogs. Cambridge, MA: Belknap Press.
Fleischacker, Samuel, 2004. A Short History of Distributive Justice. Cambridge, MA:
Harvard University Press.
Frazer, Michael, 2010. ‘The modest professor: interpretive charity and interpretive
humility in John Rawls’s Lectures on the History of Political Philosophy’,
European Journal of Political Theory 9, 218–26.
Freeden, Michael, 1996. Ideologies and Political Theory: A Conceptual Approach.
Oxford: Clarendon Press.
Freeden, Michael, 2005. ‘What should the “political” in political theory explore?’,
The Journal of Political Philosophy 13, 113–34.
Gallie, W. B., 1956. ‘Essentially contested concepts’, Proceedings of the Aristotelian
Society 56, 167–98.
Galston, William, 1995. ‘Two concepts of liberalism’, Ethics 105, 516–34.
Gardner, John, 2001. ‘Legal positivism: 5½ myths’, American Journal of
Jurisprudence 46: 199–227.
Gaus, Gerald, 2000. Political Concepts and Political Theories. Boulder, CO:
Westview Press.
Gerring, John, 1999. ‘What makes a concept good? A criterial framework for
understanding concept formation in the social sciences’, Polity 31, 357–93.
Grotius, Hugo, 2005. The Rights of War and Peace, ed. Richard Tuck. Indianapolis,
IN: Liberty Fund.
Conceptual Analysis 189

Hardimon, Michael, 2003. ‘The ordinary concept of race’, The Journal of

Philosophy 100, 437–55.
Hayek, F. A., 1960. The Constitution of Liberty. London: Routledge & Kegan Paul.
Hobbes, Thomas, 1998. On the Citizen, ed. Richard Tuck, tr.
Michael Silverthorne. Cambridge University Press.
Hobbes, Thomas, 2012, Leviathan, ed. Noel Malcolm. Oxford: Clarendon Press.
Hohfeld, Wesley Newcomb, 1923. Fundamental Legal Conceptions as Applied in
Judicial Reasoning and other Legal Essays. New Haven, CT: Yale University Press.
Honoré, Tony, 1987. Making Law Bind: Essays Legal and Philosophical. Oxford:
Clarendon Press, 161–92.
Jackson, Frank, 1998. From Metaphysics to Ethics: A Defence of Conceptual Analysis.
Oxford: Clarendon Press.
Jackson, Frank, 2001. ‘Précis of from Metaphysics to Ethics’, Philosophy and
Phenomenological Research 62: 617–24.
Joyce, Richard, 2015. ‘Moral anti-realism’, in Edward Zalta, ed., The Stanford
Encyclopedia of Philosophy, http://plato.stanford.edu/archives/sum2015/entries/m
oral-anti-realism. Accessed 20 August 2015.
Kant, Immanuel, 1997. Lectures on Ethics, ed. Peter Heath and J. B. Schneewind,
tr. Peter Heath. Cambridge University Press.
List, Christian, 2011. ‘The logical space of democracy’, Philosophy and Public
Affairs 39: 262–97.
List, Christian and Laura Valentini 2016. ‘The methodology of political theory’, in
Herman Cappelen, Tamar Szabó Gendler and John Hawthorne, eds.,
The Oxford Handbook of Philosophical Methodology. Oxford University Press,
525–53.
Locke, John, 1975. An Essay Concerning Human Understanding, ed. P. H. Nidditch.
Oxford: Clarendon Press.
Locke, John, 1988. Two Treatises of Government, ed. Peter Laslett. Cambridge
University Press.
Lovett, Frank, 2012. ‘What counts as arbitrary power?’, Journal of Political Power 5:
137–52.
Lukes, Steven, 1974. Power: A Radical View. London: Macmillan.
Margolis, Eric and Stephen Laurence, 1999. ‘Concepts and cognitive science’, in
Eric Margolis and Stephen Laurence, eds., Concepts: Core Readings. Cambridge,
MA: MIT Press, 3–81.
Margolis, Eric and Stephen Laurence, 2014. ‘Concepts’, in Edward Zalta, ed.,
The Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/archives/sp
r2014/entries/concepts. Accessed 10 July 2015.
Martinich, A. P., 2005. Philosophical Writing: An Introduction. 3rd edition. Malden,
MA: Blackwell Publishing.
Martinich, A. P., 2014. ‘Political theory and linguistic criteria in Han Feizi’s
Philosophy’, Dao 13: 379–93.
Mill, John Stuart, 1965. Collected Works of John Stuart Mill, Volumes II and III, ed.
J. M. Robson. University of Toronto Press.
Mill, John Stuart, 1969. Collected Works of John Stuart Mill, Volume X, ed.
J. M. Robson. University of Toronto Press.
190 Chapter 9

Mill, John Stuart, 1974. Collected Works of John Stuart Mill, Volumes VII and VIII,
ed. J. M. Robson. University of Toronto Press.
Mill, John Stuart, 1989. Autobiography, ed. J. M. Robson. London: Penguin.
Miller, David, 1983. ‘Linguistic philosophy and political theory’, in David Miller
and Larry Siedentop, eds., The Nature of Political Theory. Oxford: Clarendon
Press, 35–51.
Morgenthau, Hans, 1952. American Foreign Policy: A Critical Examination.
London: Methuen & Co.
Murphy, Gregory, 2002. The Big Book of Concepts. Cambridge, MA: MIT Press.
Olivecrona, Karl, 1975. ‘The term “property” in Locke’s Two Treatises of
Government’, Archiv für Rechts- und Sozialphilosophie 61, 109–15.
Olsthoorn, Johan, 2015. ‘Why justice and injustice have no place outside the
Hobbesian state’, European Journal of Political Theory 14: 19–36.
Oppenheim, Felix, 1981. Political Concepts: A Reconstruction. Oxford: Basil
Blackwell.
Outhwaite, William, 2010. Concept Formation in Social Science. London:
Routledge.
Pateman, Carole, 1988. The Sexual Contract. Cambridge: Polity Press.
Penner, J. E., 1996. ‘The “bundle of rights” picture of property’, University of Los
Angeles Law Review 43, 711–820.
Pettit, Philip, 1997. Republicanism. A Theory of Freedom and Government. Oxford:
Clarendon Press.
Pettit, Philip, 2012. On the People’s Terms. A Republican Theory and Model of
Democracy. Cambridge University Press.
Phillips, Anne, 2013. Our Bodies, Whose Property? Princeton, NJ: Princeton
University Press.
Rawls, John, 1999a. A Theory of Justice. Revised edition. Cambridge, MA: Belknap
Press.
Rawls, John, 1999b. Collected Papers, ed. Samuel Freeman. Cambridge, MA:
Harvard University Press, 47–72.
Rawls, John, 2001. Justice as Fairness: A Restatement. Cambridge, MA: Belknap
Press.
Rawls, John, 2005. Political Liberalism. Expanded edition. New York: Columbia
University Press.
Rawls, John, 2007. Lectures on the History of Political Philosophy. Cambridge, MA:
Harvard University Press.
Raz, Joseph, 1984. ‘On the nature of rights’, Mind 93: 194–214.
Russell, Bertrand, 1923. ‘Vagueness’, Australasian Journal of Psychology and
Philosophy 1, 84–92.
Sartori, Giovanni, 1970. ‘Concept misformation in comparative politics’,
American Political Science Review 64: 1033–53.
Sartori, Giovanni, 1984. Social Science Concepts: A Systematic Analysis. Beverly
Hills, CA: SAGE Publications.
Schefﬂer, Samuel, 1987. ‘Morality through thick and thin: a critical notice of
Ethics and the Limits of Philosophy’, The Philosophical Review 96, 411–34.
Conceptual Analysis 191

Schroeter, Laura, 2004. ‘The limits of conceptual analysis’, Paciﬁc Philosophical

Quarterly 85: 425–53.
Skinner, Quentin, 1996. Reason and Rhetoric in the Philosophy of Hobbes. Cambridge
University Press.
Skinner, Quentin, 2002. Visions of Politics, Volume I: Regarding Method. Cambridge
University Press.
Skorupski, John, 2010. The Domain of Reasons. Oxford University Press.
Skorupski, John, 2012. ‘Précis of The Domain of Reasons’, Philosophy and
Phenomenological Research 85: 174–84.
Stevenson, Charles Leslie, 1938. ‘Persuasive deﬁnitions’, Mind 47: 331–50.
Suárez, Francisco, 2015. Selections from Three Works, ed. Thomas Pink.
Indianapolis, IN: Liberty Fund.
Swanton, Christine, 1985. ‘On the “essential contestedness” of political concepts’,
Ethics 95, 811–27.
Tamir, Yael, 1993. Liberal Nationalism. Princeton, NJ: Princeton University Press.
Taylor, Charles, 1985. The Human Sciences: Philosophical Papers II. Cambridge
University Press.
de Tocqueville, Alexis, 2003. Democracy in America and Two Essays on America, ed.
Isaac Kramnick. London: Penguin.
Vallentyne, Peter, Hillel Steiner and Michael Otsuka, 2005. ‘Why
left-libertarianism is not incoherent, indeterminate, or irrelevant: a reply to
Fried’, Philosophy and Public Affairs 33: 201–15.
Van Parijs, Philippe, 1995. Real Freedom for All: What (If Anything) Can Justify
Capitalism? Oxford: Clarendon Press.
Von Leyden, Wolfgang, 1963. ‘On justifying inequality’, Political Studies 11:
56–70.
Waldron, Jeremy, 1988. The Right to Private Property. Oxford: Clarendon Press.
Waldron, Jeremy, 2002. ‘Is the rule of law an essentially contested concept (in
Florida)?’, Law and Philosophy 21: 137–64.
Wallis, John, 1662. Hobbius Heauton-timorumenos, or a Consideration of Mr Hobbes
His Dialogues. Oxford: A. & L. Lichﬁeld.
Wenar, Leif, 2013. ‘The nature of claim-rights’, Ethics 123, 202–29.
Williams, Bernard, 1985. Ethics and the Limits of Philosophy. Cambridge, MA:
Harvard University Press.
Wittgenstein, Ludwig, 2009. Philosophical Investigations, 4th edition, ed. and tr.
P. M. S. Hacker and Joachim Schulte. Oxford: Wiley-Blackwell.
Wolff, Jonathan, 2013. ‘Analytic political philosophy’, in The Oxford Handbook of
the History of Analytic Philosophy, ed. Michael Beaney. Oxford University Press,
795–822.
10 Positive Political Theory

Alan Hamlin

1 Introduction
This chapter is organized around two questions: what is positive political
theory, and how can it be used effectively to address research questions?
While positive political theory may apply in any substantive area of
politics, these questions will be addressed in the context of the analysis
of democratic institutions and political behaviour and, speciﬁcally, the
study of referendums, elections and voting.
The next section locates positive political theory relative to its
normative counterpart and indicates the range of positions that exist
within positive political theory. Section 3 provides an extended example
of positive political theory in the setting of the analysis of democratic
elections, so as to develop a more detailed understanding of the
component elements of any exercise in positive political theory.
Section 4 offers a ‘how-to’ guide for the development of positive political
analysis. While no such guide can guarantee success, the claim is that
detailed, systematic and explicit consideration of the processes involved
in constructing and using positive political theories can only enhance
political debate. Section 5 illustrates the ‘how-to’ guide by using its
principles as a guide to reading two journal articles.

2 What Is Positive Political Theory?

Political theory, particularly when described as political philosophy, is
often taken to be essentially normative. While it is certainly true that
normative analysis – including both the investigation of normative
principles such as justice, well-being or rights and the evaluation and
justiﬁcation of particular social and political institutions and practices –
is central to the overall ambition of political theory, normative analysis
does not exhaust political theory.
The study of politics must also be concerned with explaining and
understanding social and political institutions and practices and the

192
Positive Political Theory 193

political behaviour of individuals within those institutions and practices.

Indeed, the tasks of explaining and understanding might be argued to be
logically prior to the tasks of evaluating and justifying; if we cannot say (to
some degree of approximation) how a particular institution operates, how
can we evaluate that institution? It might be logically possible to offer
a fully deontic justification of an institution that does not depend on an
evaluation of its workings and the outcomes arising, but most normative
approaches would place at least some weight on those outcomes, or the
behaviour arising within the institution.
Explanation and understanding cannot simply be a matter of
description or direct empirical observation. This is obvious when we
seek to understand an institution that does not currently exist (perhaps
to consider bringing it into existence), but it is equally true where an
institution exists. However detailed our description of an institution’s
operation, we cannot fully explain or understand it, as our observations
are limited to the particular circumstances that we happen to have
experienced. A full explanation and understanding would also offer
a counterfactual account of the institution’s operation in other
circumstances.
Positive political theory is that part of political theory that attempts to
fill the gap between description and normative analysis, providing us with
explanations of political phenomena and behaviour that are both crucial
to our understanding of politics and essential to our normative discussion.
Whenever we offer an account of any political event or institution, we are
engaging in positive political theory. The ubiquity of positive political
theory sometimes renders it invisible, in much the same way that
Monsieur Jourdain fails to see that he normally speaks in prose (in
Moliere’s ‘Le Bourgeois Gentilhomme’). We can hardly engage in any
political discussion without invoking elements of some positive political
theory, but we often fail to recognize that fact, or the implications and
limitations of the particular theory that we are invoking. A main theme of
this chapter is that our discussions would be improved if the underlying
positive political theory component were more explicit.
The use of positive political theory should not be confused with
a commitment to positivism. This is not the place to engage with the
wider debate on positivism (Gordon 1993; Kincaid 1998), but it is worth
pointing out that one can take any of a variety of non-positivist views in
the philosophy of social science without undermining the significance of
what I refer to as positive political theory. Theories may be Marxist (or
post-Marxist), structuralist (or post-structuralist), feminist (or post-
feminist) or embody any of a wide variety of further commitments but
still form part of positive political theory in the sense that I intend it.
194 Chapter 10

Similarly, theories may draw on anthropology, economics, psychology,

sociology or other disciplines while maintaining their essentially political
character.
If positive political theory is not necessarily ‘positivist’, we should also
note that it need not refrain from all use of normative terms and ideas.
Many positive models in politics will involve assumptions about the
motivations of political agents, and many of the motivations studied
are ‘normative’ in character: we might, for example, study the behaviour
of individuals motivated by considerations of ‘justice’, or by considera-
tions derived from a broader morality. While the motivations under
consideration are normative, our study can still be positive if our focus
is on understanding the behaviour of individuals with the specified
motivation rather than evaluating or justifying that motivation. In this
way, elements of our study can be recognized as normative in them-
selves, but still taken as the object of positive study: a piece of positive
political theory can include reference to normative terms and concepts
provided that the reference is of an appropriate kind.
Having identified positive political theory in this very expansive way,
I should note that the phrase ‘positive political theory’ is often used more
narrowly, sometimes to mean ‘formal political theory’, sometimes to
mean ‘rational choice political theory’ and sometimes to mean ‘game-
theoretic approaches to politics’. I will comment on each of these usages.
Formal political theory identifies that sub-class of political theory that
is expressed in the style of theorems and lemmas using the tools of
mathematics or formal logic. The defining feature of formal political
theory is simply its mathematical or logical formality rather than the
topic that the theory addresses, the particular nature of the assumptions
made within the theory, or whether the theory is positive or normative in
character. Formal political theory is often, but not always, linked to
detailed statistical modelling (Riker and Ordeshook 1973; Forbes 2004).
Rational choice political theory identifies the sub-class of political
theory that takes as foundational the assumption that individuals act
rationally. Within this class we may find both formal and informal
theories, and both positive and normative discussion, although it is
certainly the case that much rational choice political theory is relatively
formal and positive in its orientation (Satz and Ferejohn 1994; Shepsle
and Bonchek 1997; Austen-Smith and Banks 2000; Hindmoor 2006).
For critical discussion of the rational choice approach, see Green and
Shapiro (1994), Friedman (1996) and Hindmoor (2011). For more
details, see Kogelmann and Gaus’s chapter in this volume.
Game-theoretic political theory, also discussed in Kogelmann and
Gaus’s chapter in this volume, is a subset of rational choice political
Positive Political Theory 195

theory. Game-theoretic accounts not only take individuals to be

rational in a particular sense, but the situation under study is taken to
constitute a ‘game’ in the sense that it is the strategic interaction
between rational individuals that is emphasized (Ordeshook 1986;
Austen-Smith and Banks 2005). While it is possible to discuss game-
theoretic political theory in a relatively informal way, game theoretic
analysis is built on formal foundations. But game-theoretic approaches
to politics are not the only possible intersection of formal methods and
rational choice: it is possible to identify formal, rational choice models
that are not game theoretic.
If formal, rational choice and game-theoretic approaches are only
subsets of positive political theory, why are they sometimes claimed to
occupy the whole of the territory? In part this is simply a matter of
prominence: there can be little doubt that the literature that uses the
language of positive political theory most explicitly is closely associated
with approaches that combine formality, rationality and game theory in
various combinations. But, to revert to the analogy with prose, just
because some prose speakers proclaim that they speak in prose, it does
not follow that they are the only prose speakers. Informal positive political
theory, by the very fact that it is informal, tends to be much less self-aware
than its formal cousin, but whenever a writer makes any claim about the
behaviour of individuals or social groups in political settings, or the
characteristics of a particular political institution, or makes a prediction
about political events, he or she will normally be relying on some under-
standing of underlying forces or patterns of causality that merits the use of
the term ‘positive political theory’. Of course, many such ‘theories’ are
largely implicit, and some may be such that any attempt to make them
explicit would reveal them to be little more than prejudice or opinion. But
moving from the implicit to the explicit and improving theories (however
we might deﬁne ‘improving’) is not necessarily the same as formalizing
those theories (in the sense of presenting them in mathematical form) or
rendering them within the framework of rational choice or game theory.
A positive political theory may be useful, revealing and insightful while
being informal and making no signiﬁcant reference to rational choice; just
as a formal, rational choice or game-theoretic theory may also be useful,
revealing and insightful.
So, what are the essential elements of a positive political theory? Here
there is no clear, universally accepted answer, but it seems relatively
uncontentious to begin with the idea of a model. Initially we might
think of a model as a limited representation of some element of reality.
It is important that the representation of reality is limited, since the reason
we create models is that reality is too complex to be studied in its raw
196 Chapter 10

form. But it is also important to consider how a model’s representation of

reality is limited. There are three key elements here: abstraction,
simplification and idealization. Abstraction is the idea that we manage
some aspects of the complexity of the real world by ignoring them; that is,
by leaving them out of the model altogether. Clearly, we would like to
abstract from those aspects of reality that we think unimportant, but often
we may need to abstract from some potentially important areas in order to
focus our attention on others. Simplification is the idea that, even when
we include an aspect of reality in our model we will typically need to
include only some of its features. Again, we would like to focus on the
most salient or significant features, but we will often have to sacrifice
features of interest in order to focus our study. Idealization is the idea that
in modelling some feature of the real world we may need to represent that
feature in a rather stylized or ‘pure’ form, rather than the messy and
complicated forms exhibited in the real world.
A simple illustrative example of a model from outside of the social
sciences may help. Consider the London underground map. This is
a model. It abstracts from many aspects of real-world London,
completely ignoring streets and features such as parks or buildings in
order to focus on the layout of the underground network. It simplifies
the depiction of the underground network, so that, for example, the map
is not to scale and does not depict the real geographical relationship
between underground stations. It also idealizes the network to present
a graphically striking image relying on colours and design features that do
not correspond to the underlying reality.
This example suggests an important point: models are created for
a purpose; they are good models to the extent that they serve their purpose
well. In particular, good, useful models do not need to be ‘realistic’ in any
general or complete sense; indeed most good models will abstract from,
simplify and idealize reality to such an extent that they are clearly
‘unrealistic’. Of course, a good model will retain some connection to the
real world; but that connection may be highly stylized, so that the
relationship between the model and the real world is less like a detailed
photograph and more like a caricature that captures one or two key
features of reality in a simplified (even exaggerated) form while ignoring
everything else.
The general point here is that a model is to be judged by its usefulness
rather than by any direct appeal to its realism (or the realism of its
assumptions). And this in turn suggests that one might want many
different models of essentially the same piece of reality, with each model
aiming to capture a different aspect of that reality: just as one might want
many different maps of London in addition to the underground map,
Positive Political Theory 197

each serving a rather different function, so one might want a variety of

models of a political election, each focusing on a different aspect of the
complex whole. Rather than these different models being rivals, they may
complement each other, so that each model contributes to our more
general understanding.
In moving from models such as the London underground map to
models and theories in the social sciences, we need to address two
complications: one concerned with a further aspect of the make-up of
political models, the other with the idea of ‘usefulness’.
So far, a model has been identified as an abstract, simplified and
idealized representation of a part of reality. But most models in politics
involve another feature, one not shared by the London underground map.
This is some animating idea that adds a structure of causality to the
model. It is at this point that a model becomes the carrier of a particular
theory (Elster 1989; Van Evera 1997). For example, in the study of voting
behaviour, an animating idea might be that a particular individual’s
voting behaviour is explained in part by the social group of which he or
she is a member and in part by the extent of that individual’s interactions
with members of other social groups (Zuckerman et al. 1994; Abrams
et al. 2011). A completely different idea of voters’ motivation is that
voters vote for the candidate they find most physically attractive
(Berggren et al. 2009). Many other ideas with at least some plausibility
are possible, but, whichever idea is selected, the same principles relate to
the specification of an animating idea or causal theory as relate to the basic
construction of the model: principles of abstraction, simplification and
idealization.
The mere fact that just one or two of the wide range of possible
motivational or causal ideas are selected in any particular study is
sufficient to demonstrate the principle of abstraction in this context.
The motivations of real individuals are hugely complex, both in the
sense that any single individual is likely to display a wide range of
different motivations and in the sense that different individuals are
likely to display different motivations; but a theory must place some
limits on the degree of complexity it admits, and most theories will
focus attention on a very small subset of potential motivating or causal
influences. Similarly, theories will generally need to simplify the form
of the particular motivations under consideration, perhaps by
restricting the degree of variation across individuals, or by imposing
a particular and somewhat arbitrary definition on what features are
considered relevant. Finally, theories will typically idealize the
motivation of the individuals by taking extreme or pure cases, which
might entail making unrealistic assumptions about such matters as the
198 Chapter 10

extent to which individuals have access to relevant information, or the

extent to which they are consistent in their behaviour.
A basic model provides a setting in which we can isolate key aspects of
reality, but without an animating idea or theory such a model is passive: it
does not generate any particular understanding of the ways in which these
key aspects interact to produce outcomes. This should not be taken to
imply that such basic models are not valuable or useful. A basic model will
be useful if it frames and addresses a research question in a way that is
helpful: just as the underground map can help one to navigate across
London, so a basic political model can help one to navigate the literature
on a particular political question. But it is the addition of an animating
idea or causal theory that transforms the model into an active tool for
political investigation.
The second complication relates to the idea of ‘usefulness’. While
the idea of the usefulness of a basic model such as a map is easy to
understand, it is more difficult to be precise about the ‘usefulness’ of
a more animated theory in politics. The general ambition of most
models and theories is to contribute to our understanding of some
political phenomenon; but how can we tell if a model is indeed useful
in this way, and how can we combine the insights offered by different
models? Part, but only part, of the answer lies in the relationship
between theoretical models and empirical work (whether quantitative
or qualitative).
One way in which a theory may be useful is in its ability to explain or
predict observed empirical patterns. For example, a theory that sets out to
explain differences between different voting systems might be expected to
cast at least some light on the patterns of results thrown up by those voting
systems in the real world, and perhaps even make some predictions about
future results. While this seems reasonable, it may also be difficult to
achieve in practice. It is unlikely that direct empirical observation of two
or more voting systems operating in otherwise identical environments is
possible, so that empirical data will always be at least somewhat difficult
to interpret. In some cases relevant data simply may not exist, but the
issue goes rather deeper than this. Until we identify relevant models and
theories, we do not even know what data may be relevant, and so what
data to collect. If a theory tells us that some factor X may be important in
explaining a political phenomenon, then this may persuade us to collect
data on X so as to be able to ‘test’ the theory or at least to investigate the
relationship in more detail. But notice that the data are already ‘theory-
laden’ in the sense that we are sensitized to that particular view of the
world because of the particular theory adopted. Had we adopted
a different theory, one positing a relationship between Y and the political
Positive Political Theory 199

phenomenon in question, we might have gathered other data and reached

other conclusions. We must avoid the trap of thinking that there is some
pre-theoretic stock of ‘data’ that speaks for itself in guiding our choice of
theories.
Even if this trap is avoided, empirical relevance is not the only sort of
‘usefulness’ that a model might achieve. A model will often serve to focus
attention on the linkages between research questions and the way in
which the exploration of an issue can be extended. In this way, a model
can inﬂuence the course of development of a literature, suggesting
connections and further developments that might not have been
recognized in the absence of the model. If in constructing our basic
model we recognize explicitly that we are abstracting from some poten-
tially relevant factor, this will focus attention on the question of extending
the model to incorporate that factor in more developed models. Different
models will suggest different developmental paths, and this is another way
in which a variety of modelling approaches can be complementary.
Models are not just static objects; models develop over time with many
authors contributing to the model in different ways. Each development
throws up new challenges and criticisms, and these will in turn provoke
further work both within the same model and in other models as research-
ers react to each other’s arguments. In this way, the variety of models
employed by political analysts may be thought of as a network of pathways
that criss-cross the territory of politics. Each pathway may claim
something distinctive, but it is the growth of the network that reﬂects
the real range and depth of political research.

3 An Extended Example
At this point it is useful to provide an example, to illustrate the various
points made. The example begins with an extremely simple model of
a referendum in the broadly rational choice tradition, and shows some of
the ways in which that model has developed over time. While our focus
here is on the positive theory, note that the ultimate goal here, as is almost
always the case, is normative – we study referendums and elections not
just to understand them, but so as to be able to improve the design and
performance of our political institutions. Yet the normative concern for
improvement requires a relatively detailed positive understanding.
In keeping with the ideas of abstraction, simpliﬁcation and idealiza-
tion, we will begin by identifying the minimal necessary ingredients for
a positive model of a referendum: a set of voters, some issue over
which the voters disagree, two alternative policy positions with respect
to that issue, and a voting rule. A referendum, in this simple model
200 Chapter 10

Vote for X P Vote for Y

L X Y R

Figure 10.1: The basic Downsian model

world, is simply the choice of one of the policy positions by the set of
voters acting through the voting rule. To be specific, and to idealize
features of any real-world referendum, assume that the voting rule is
simple majority voting (note that with just two alternatives, almost all
plausible voting rules converge on simple majority voting) and that the
issue at stake can be described as choosing a point along a left-right
spectrum. Left-right here does not need to carry any particular political
significance; it might, for example, be the level of public spending on
a particular activity, or the tax rate to impose in a particular context.
Assume also that each individual voter identifies an ideal point on the
spectrum at issue and would like the outcome to be as close as possible
to that ideal point. This adds an element of motivation to the
individuals in the model and is what makes this a model in the broadly
rational choice tradition: we assume that each individual acts in the
way that she believes will contribute to bringing about the best
available outcome seen from her own perspective. This does not
amount to an assumption of self-interest. The individual may choose
her ideal point because she believes it to be in the public interest, or
because she believes it is morally best, or for any other reason. All that
‘rationality’ requires here is that the individual identifies a relevant
ideal point and acts so as to bring about the closest possible approx-
imation to that ideal. This is essentially the model of democratic
decision-making introduced by Downs (1957) and may be illustrated
diagrammatically.
In Figure 10.1, the L-R line represents the policy issue at stake, with
points along the line representing different possible positions that might
be chosen. X and Y are the two specific policy positions that are ‘candi-
dates’ in this referendum, and the voters may be thought of as spread
along the L-R line with each voter positioned at his or her ideal point.
If we define point P to be the point halfway between X and Y, and we
assume that everyone votes (note the idealization here), it should be clear
that all voters whose ideal points lie to the left of P will vote for X and all
Positive Political Theory 201

those to the right of P will vote for Y. Given the simple majority voting
rule, X will win if a majority of voters lie to the left of P, Y will win if
a majority of voters lie to the right of P.
So far this model does nothing more than articulate the idea of majority
voting in a simple setting. To animate the model further we might add
another element. Consider the strategic choice of X and Y on the
assumption that these policy positions are chosen by political parties
whose choice is guided by the desire to maximize the probability of their
proposal winning the referendum. Note that we are here introducing
a second element of motivation, and again we are making that motivation
as simple and stark as possible, even though this may be unrealistic.
If, in the initial position depicted by Figure 10.1, X would win the
election, the political party controlling Y would face an incentive to shift
Y leftward. By doing so, the position of point P will move to the left, more
voters will support Y and fewer will support X. But similarly, the party
controlling X will face an incentive to move rightward, so increasing its
vote, and reducing its rivals. This suggests that the two party platforms
will converge under competitive pressure; but where might this process
stop? One aspect of the answer is that in the absence of any further
argument, there is nothing to stop the two platforms converging to
a single point, so that we might expect the two parties to offer the same
policy.
But this is only half the answer. Imagine that both parties offer policy
Y in Figure 10.1, and that more than half of the voters’ ideal points lie to
the left of Y. It is straightforward to argue that each party now faces an
incentive to move leftward. If either party succeeds in positioning itself
just to the left of its rival, it will win the referendum. But if both parties
face this same incentive, we might expect both to react (one of the
simplifying features of the model is that the two parties are essentially
identical). Similarly, if both parties chose a policy platform such that the
majority of voters’ ideal points were to the right of that policy, then both
would face an incentive to move rightward. So the model tells us not
only that the two parties will converge on the same policy, but that there
is a unique policy point at which the two parties will settle: the policy
that is the ideal point of the median voter, the point at which exactly
half of the voters’ ideal points lie to the left and half lie to the right. This
is the ‘Median Voter Theorem’ that says that in a two-candidate contest
of the type described, both candidates will offer a policy platform aimed
at the median voter’s ideal point.
This is a very simple model, and its simplicity generates both a clarity of
argument and a range of suggestions for further work. And these two
things are closely related. It is precisely because we can see the mechanics
202 Chapter 10

of the model clearly, and understand the forces at work, that we can
pinpoint important limitations of the model and identify further research
questions. A series of questions interrogates the robustness of the model.
What would be the impact of relaxing the assumption that all citizens
vote? What would be the impact of introducing a third political party?
What would be the impact of assuming that political parties had ‘ideal
policies’ of their own that tempered their motivation to win the vote at all
costs? What would be the impact of allowing the vote to operate on more
than one political issue (so that it becomes a model of an election rather
than a single-issue referendum)? How might we compare different voting
system in this framework? How might the model be reformulated to
capture the idea of electing representatives rather than making direct
policy choices?
Some of these questions are relatively simple to address, others require
considerable detailed work, but all of these questions, and many more,
have been explored in the literature that has developed since Downs
(reviewed in Mueller 2003). For example, the issue of abstention opens
up the question of identifying the factors inﬂuencing turnout (Aldrich
1993; Blais 2000; Bendor et al. 2003). One possibility is that voters
abstain when they are essentially indifferent across the alternatives on
offer; another possibility is that voters abstain when the policy platforms
are too far from their own ideal points (voter alienation). These
possibilities have different implications. If political parties converge on
identical policy platforms, and individual citizens abstain when they are
essentially indifferent between the platforms on offer, the theory will
generate predictions of low turnout. One branch of the literature intro-
duces the idea that individuals might participate in the vote out of a sense
of civic duty, even if they still vote for whichever platform is closer to their
ideal (Riker and Ordeshook 1968; Ferejohn and Fiorina 1974). This
opens up the idea that the factors that drive turnout may be rather
different from the factors that drive which option to vote for.
A second example relates to animating individual voters. The initial
idea in play here is just that individuals vote instrumentally to bring about
whichever outcome they see as ‘best’ regardless of exactly how they deﬁne
‘best’. But perhaps voting behaviour might be modelled differently: in
terms of habitual voting, retrospective voting or expressive voting.
The idea of habitual voting could be incorporated by assuming that an
individual’s vote in any particular election is partly determined by his or
her votes at earlier elections (Fowler 2006). This might allow parties to
pursue non-convergent platforms if they felt that their habitual vote was
secure. The idea of retrospective voting is that voters may be backward-
looking when choosing how to vote, rewarding good (or punishing bad)
Positive Political Theory 203

behaviour by parties in previous periods, rather than focusing on their

proposals for the future. This idea engages with the idea of an
incumbency effect (Fiorina 1981; Krehbiel and Wright 1983; Ferejohn
1986; Fiorina et al. 2003). The idea of expressive voting is that individuals
may vote to express some aspect of their identity, beliefs or personality
rather than voting to bring about a particular outcome, and that this is
particularly likely in large elections where an individual is extremely
unlikely to be instrumentally signiﬁcant in determining the outcome of
the election (Brennan and Lomasky 1993; Brennan and Hamlin 1998;
Hamlin and Jennings 2011).
A third example focuses on the role of political parties (Duverger 1965;
Panebianco 1988; Strøm 1990). In the original model, parties are
sketched as independent agents who seek only to win elections, and this
immediately raises questions relating to the structural relationship
between political parties and their members who are themselves also
citizens and voters, and further questions relating to the mechanisms
and processes by which parties choose their policy platforms.
Extensions to the model develop a number of aspects of political parties
including: allowing individual citizens to stand as candidates (Besley and
Coate 1997); viewing political parties as operating to extend political
credibility over time (Brennan and Kliemt 1994); discussing the decision
of individuals to join political parties (Hamlin and Jennings 2004); and
discussing the internal choice of party leaders and the relationship
between leaders and policy platforms (Ware 1992; Hamlin and Jennings
2007).
These examples serve to illustrate the genealogy of models: the way in
which models and theories develop over generations of academic debate;
and the diversity of the resultant ‘family tree’. Understanding how
a particular model ﬁts into the broader landscape of such family trees is
an important part of appreciating both the richness and the limitations of
that family of models.

4 How to Do Positive Political Theory

While this extended example has provided a view of some of the general
issues that arise in building, developing, understanding and using
a positive political model/theory, it is now time to make explicit some of
the lessons that have been implicit in the past two sections.
In practice it will often be the case that a model or theory is adapted
from the existing literature (as suggested by our extended example) rather
than designed from scratch. But whether you are attempting to build
a model from scratch, adapt a model from the literature, or simply
204 Chapter 10

Table 10.1: Essential elements of a model of a referendum

Essential Structures Inessential Structures

Citizens/Voters Any group structure of individual voters

An issue Multiple issues
Two policy options/political parties Multiple options/candidates
Voting rule Other institutional features (representation,
repeat elections, etc.)
Etc.

understand a model in the literature, it is sensible to approach the exercise

in much the same way. The remainder of this section will identify a series
of steps that provide a guide to positive modelling.
The starting point should be to identify your research area and
basic research question in their simplest possible forms. Begin by
simply listing the major structural features of the research area that you
believe are essential in any model that could possibly address your
research question, and separately list those that might be excluded at
this initial stage even if their inclusion might seem desirable. The idea is
to sketch the ingredients for the simplest possible model at this
stage. In terms of our example of the Downsian model, the essential
elements of the model are listed in Table 10.1, as are at least some of the
more obvious structural features that might be seen as optional extras.
The structures that you see as essential will depend on the precise focus
of your research question. If the intention is to study the impact of
campaigning on referendum outcomes, then the simplest speciﬁcation
of the essential structures to include in the model will be rather more
complex than that shown here since you will need to include at least some
features of campaigning. The point, however, is to arrive at the simplest
list of ingredients that offer the possibility of modelling the area of your
concern.
The next step is to sketch the basic relationships among the
identiﬁed ingredients of the model. For example, many models in
the area of electoral politics include both ‘citizens/voters’ and ‘political
parties’ as structural features, but models will differ in the focus they wish
to place on the relationship between these elements of the model. In some
cases we may wish to simplify and idealize our view of political parties.
In other cases, it may be essential to the intended purpose of the model to
consider the internal structure of political parties and the way in which
their policy platforms emerge (for a variety of approaches, see the papers
Positive Political Theory 205

CITIZEN/ PARTY VOTING

ISSUE
VOTERS PLATFORMS RULE

OUTCOME

Figure 10.2: Schematic outline of basic Downsian model

collected in Müller and Strøm 1999). Clearly, the structure of the

relationship in the model will be quite different in the two cases. It may
be useful to transform the simple list of the type illustrated by Table 10.1
into a diagram of the form of Figure 10.2, which shows the basic structure
of the relationships that are key to the model in its simplest form.
The items along the top row are the basic inputs of the model, as specified
in Table 10.1. The only item that is actually determined within the model
is the outcome of the referendum, and that is seen as influenced by all of
the independently specified features of the model, as indicated by the
arrows. As is clear from Figure 10.2, the structure of the basic Downsian
model is particularly simple, with a direct relationship between each of the
basic features and the single outcome, and with no complicating features
such as interactions among the features, or feedback from one part of the
model to another.
A slightly more complicated version of the model, allowing citizens to
abstain depending on the voting rule and their view of the platforms
adopted by the parties, might be sketched as in Figure 10.3. Note that
party platforms and the voting rule now have two effects on the final
outcome: a direct effect, as in the simple model, and an indirect effect via
citizens’ decisions to vote or abstain.
At this stage we have a basic, passive model, comparable to the London
underground map. It offers a simple guide to the research area under
investigation that allows us to consider the various possible linkages
between the identified features. It also suggests ways in which we might
extend and complicate the model to include features that we might
believe to be important (if not absolutely essential).
At this stage it is worthwhile to reflect on your proposed approach
to further study and the purposes that you want your model to
serve. One purpose common to most pieces of research is to provide
206 Chapter 10

PARTY VOTING
ISSUES CITIZENS
PLATFORMS RULE

VOTERS

OUTCOME

Figure 10.3: Schematic outline of Downsian model with possible

abstention

a structure for reviewing the literature. A basic model of the type

constructed so far can be valuable as a way of thinking about and compar-
ing alternative accounts of referenda in the literature or in general political
debate. Each account should be capable of being analysed in terms of our
basic schematic structure, by means of a series of simple questions. How
does the account specify each of the ingredients of the model? What
assumptions are made, explicitly or implicitly, about the set of voters, or
the number and nature of political parties and the way in which they set
their policy platforms? What are the properties of the outcome of the
referendum in the account under consideration, and how do these
properties follow from the assumptions made? But beyond framing
a literature review, the purposes of a model will vary from case to case.
For example, if the intended study is empirical and quantitative, the
model will provide the ﬁrst step towards specifying the key variables and
data requirements. If the intended study is qualitative and interview-
based, the model will suggest key questions that should be asked and
the nature of the relationships that should be probed. If the intended
study is to develop a normative discussion of some aspect of institutional
design, the model will suggest the key connections between behaviour
and institutional arrangements that will need to feature in that normative
account. But whatever the original intention, the design of the underlying
Positive Political Theory 207

positive model will play a key role, ensuring that you move back and
forth between considering the features of the basic positive model
and the intended purpose of the proposed research in an iterative
and flexible process.
The next step is to animate the model and be explicit about
motivational and causal aspects of the model. In the case of the
basic Downsian model we noted two such aspects: the assumed
motivation of the voters in deciding how to vote and the assumed
motivation of the political parties in setting their platforms. Clearly
many other motivational assumptions are possible even within this very
simple structure. At this early stage, motivations and causal forces should
be as explicit and as simple as possible. One way of thinking about this is
to consider each of the arrows in a figure such as Figure 10.2 or 10.3, and
draft a clear statement of the nature of the relationship represented by that
arrow. The exercise of drafting such an explicit statement will almost
always bring three points to the fore. The first is that apparently simple
statements of a causal or motivational relationship often leave open
a considerable range of interpretation, so that detailed thought is required
to construct a clear, explicit statement of the relationship you have in
mind. The second is the requirement for some degree of coherence or
consistency as between the various relationships within the model.
The third is that there are almost always many quite different ways of
identifying a particular relationship, each of which carries at least some
degree of plausibility, so that many different, but related, theoretical
models could be constructed.
The first point is an example of the role of theory construction in
the process of conceptual clarification. By setting yourself the task of
thinking explicitly about the relationship between two parts of your
model and specifying that relationship as clearly and concisely as
possible, you should think carefully and clearly about the
concepts involved. Constructing and understanding your model,
albeit a simplified, abstract and idealized model, involves considerable
investment in the ideas and concepts that are basic to your research.
At this point the style and formality of a theory or model may come
under scrutiny. As your understanding of the model and its
component parts deepens, you will need to find a means of
communicating the nature of the model to others that reflects
its structure and the detailed specification that you have
decided on. There is no uniform answer to how best to present
a theory/model, but you should at least be aware of a variety of
options and make a conscious choice of presentational strategy
that fits with the overall research plan. It may be that an entirely
208 Chapter 10

textual approach is appropriate, in other cases the use of a diagram or

flow chart or other device may be helpful, and in still other cases
a greater degree of formalism may be appropriate. But whatever
style is adopted, the underlying aim is clarity of communication,
discussion and analysis.
The second point can be read at a variety of levels. At one level we
might ask whether the motivational elements of the theory/model
fit together coherently. In the simple Downsian model this is easily
achieved since the only two motivational elements of the model relate to
two distinct groups of agents (individual voters and political parties). But
in other cases there may be some tension. For example, imagine that we
are constructing a model that involves individuals voting and those same
individuals making a decision on where to live, as might be the case if we
are interested in differences between political constituencies or regions
(Tiebout 1956). And imagine that we specify the motivations guiding the
location decision in terms of maximizing some notion of self-interest, but
at the same time specify the motivations guiding the voting decision in
terms of some notion of the public interest. This raises the question of
how we are conceiving of the individual overall. The tension between self-
interest and public interest should force us to think about the underlying
model of the individual and her decision making (Buchanan 1984). Such
tension is potentially creative, can help to focus on issues that were not
immediately apparent and can help to develop interesting ideas. A benefit
of explicit and detailed modelling/theorizing is that it can both reveal such
tensions and help to ensure that they contribute positively to the overall
analysis.
The next step is to consider a range of variations on the theme of
the basic model. Recognition that there is no uniquely privileged theory/
model in relation to any particular research issue, and that there are many
potential models that can claim to be of significant interest, may at first
sight sound like a problem. But a theme of this chapter has been that
a ‘good’ model is a useful model, and there are many ways to be useful;
but just because there are many good/useful models in any area of politics
this does not imply that all models are good/useful.
Once a theory/model has been constructed, that theory/model should
be reconsidered by explicitly viewing variations. Here the general rule is
simple enough: vary one aspect of the theory/model at a time in
order to consider the impact on the model overall. If a particular
variation makes little or no difference to the model, this provides a basis
for a simple generalization of the model; if, on the other hand, a variation
does make a significant difference to the overall model, you have found
Positive Political Theory 209

a potentially interesting feature that can be incorporated into your

analysis.
Performing such a ‘sensitivity analysis’ – exploring the sensitivity of
a model to changes in the specification of its component parts – deepens
understanding of the model and helps to identify which of the
assumptions embedded in the model are merely simplifying and which
are vital to the model’s conclusions. In the simple Downsian model, it
might have been thought that the nature of the distribution of voters’ ideal
points along the L-R spectrum would play an important role in determin-
ing the way in which political parties would choose their platforms, and
whether those platforms would converge. But the sensitivity analysis
performed by Downs revealed that the precise distribution of voters
makes no difference at all, so that he did not need to make any assumption
about it. All that matters is that the distribution has a median, and this is
true of all well-defined distributions in a single dimension. In this way the
sensitivity analysis does two things. It generalizes the model (rather than
being a model that applies only when the distribution of voters is of some
particular form, it is a model that applies to all one-dimensional
distributions). And it points to a key feature, the existence of a median
that provides the basis for further investigation, since distributions in
more than one dimension (i.e. models with more than one political
issue being decided) do not generally have a median.
It is inherent in the nature of a model that no useful model can
incorporate all potentially relevant aspects of reality. Thus the fact
that a theory/model has limitations, in the sense that it is sensitive to
some changes in its basic assumptions, is inevitable and is not in itself
a criticism of the theory/model. Explicitly recognize the
limitations of your model/theory – it can only enhance its
usefulness. Sensitivity analysis can help to improve the theory/
model by pointing to the areas where extensions to the model pro-
mise significant results.
A final step is to consider zooming in and zooming out on your
proposed model. A particular way of varying a theory/model in order
to understand its properties and develop them to greatest advantage
might be termed zooming in or zooming out. Any particular model
operates at a particular level of detail: it might be a relatively ‘macro’
model focused on the big picture, it might be a relatively ‘micro’
model focused on specific details within that picture, or it might
operate in the middle ground of a ‘meso’ model. But whatever the
level of the model, it can be very useful to explore what, if
anything, the model says in relation to other levels. We can
test out the general plausibility of the model and, as with other
210 Chapter 10

forms of sensitivity analysis, distinguish the important from the trivial,

by taking a micro model and zooming out to the macro level, asking
questions like ‘what would be the implications if the assumptions of
this model where applied more generally?’; or by taking a macro
model and zooming in to more speciﬁc details, asking questions like
‘what would my general model say about some particular case study?’.

5 Application
This approach to developing and using positive political theory is
intended to provide a flexible structure for thinking about a wide variety
of political questions and issues. Clearly it needs to be fine-tuned and
adapted to fit any specific purpose. While it has been presented as a way of
approaching research, it can also provide a way of reading the literature
and, to illustrate its use in this way, this section considers two articles in
the general area of democracy and voting that are published in the same
issue of The British Journal of Political Science.
Before turning to these articles, one important point should be
stressed. There is a crucial difference between the work done in
a research project and the report of that work in a final document.
In some contexts (a PhD thesis or monograph) it may be both
appropriate and important for the final document to explicitly display
almost all of the process of research – the consideration of variations
on a theme, the zooming in and zooming out and so on. In other
contexts, such as a journal article, the final document will typically
focus on a relatively small part of the overall research, presenting just
the key message and argument. Much of the research process will be
implicit in the way the article refers and relates to the literature and in
the way in which claims and results are framed.
With this thought in mind, consider two recent articles that relate to
models of democratic behaviour. One is theoretical and normative, while
the other is empirical but, despite this difference, I will point up the extent
to which both depend on underlying positive political theory and how the
approach outlined earlier helps us to read these articles and understand
the broader research programmes of which they form part.
The more theoretical article, by Annabelle Lever (2010a), is focused on
the normative question ‘should voting be compulsory?’ (See also Hill
2010; Lever 2010b.) Lever is clear from the outset that her approach
requires consideration of the differential impacts of compulsory and
voluntary voting systems, so that a positive account of each system is
a required input to the normative debate. Lever identifies a general chain
of argument in support of compulsory voting that proceeds through
Positive Political Theory 211

a series of steps: voluntary voting results in low turnout, low turnout

normally implies unequal turnout across social groups, unequal turnout
reinforces social disadvantage, therefore compulsory voting addresses
social disadvantage. Furthermore, compulsory voting carries a range of
further benefits over and above those that depend on increased turnout.
Finally, compulsory voting carries no significant costs and does not
threaten liberty, in large part because not voting is seen as a form of free-
riding. She then criticizes each link in this chain of argument and develops
her own sceptical view of compulsory voting.
While the explicit focus of the article is normative, there are clear
indications of the underlying positive structure and equally clear, if impli-
cit, indications of the use of the principles outlined previously. The article
does not explicitly lay out the underlying model of the behaviour of voters,
the relationship between the decision to vote and other socio-political
variables (particularly those associated with disadvantage), or the signifi-
cance of the act of voting relative to other political acts in representative
democracies, but it is clear that these factors are crucial parts of the
argument, and equally clear that they have to be viewed as parts of an
overall, integrated model of political behaviour. This reliance on an
underlying model can be read into the references cited, but is also visible
in the argument developed. For example, the discussion of the view of
non-voting as a free-rider problem is necessarily embedded in a game-
theoretic understanding of electoral behaviour. Similarly, the discussion
of the possibility that the argument for compulsory voting might depend
in part on other features of the voting system (e.g. whether a first-past-
the-post or a proportional representation system is in place) provides an
example of recognizing the additional factors that might be included in
any model and thinking in terms of variations on a theme. As already
mentioned, an article-length discussion will not usually display all the
features of the underlying analysis, but in reading such an article we
should make the effort to consider the range of sensitivity analysis and
the forms of zooming in and out in relation to the model that might be
appropriate.
Any comprehensive answer to the question ‘should voting be
compulsory?’ necessarily involves the development of a range of positive
political models alongside an understanding of a range of normative
considerations. Such a task identifies a major research programme.
Lever’s article contributes to that programme, but locating and
evaluating that contribution require much more than an internal reading
of the article itself. The understanding of positive political theory devel-
oped earlier provides both a way to add value to the reading of such an
article, and a way to think about the wider research programme.
212 Chapter 10

In the second, more empirical article, Jeffrey Timmons (2010) also

directly addresses an explicitly stated research question: ‘does democracy
reduce economic inequality?’ While much of the focus is empirical in the
sense that Timmons is concerned with issues such as the country samples
utilized in various studies, the time period studied and the empirical
methodology employed, there is also a significant focus on identifying
arguments that lead us to expect that democracy may be causally related
to economic inequality, and these arise from positive political models. For
example, the median voter model together with an empirical observation
to the effect that the distribution of income or wealth is skewed (so that
the median voter’s income/wealth lies below the mean income/wealth
level) generates an argument that a programme of redistribution from
rich to poor can be expected to find majority support. Similarly, one
might argue that models that stress that democratic support for expendi-
ture on public goods such as education or health might be expected to
benefit the relatively poor more than the relatively rich.
But, as Timmons points out, the links between these positive models
and the available empirical data are relatively weak. The median voter
model suggests that post-tax and benefit income may be more equally
distributed than pre-tax and benefit income in democracies, and that
democracies may be expected to engage in redistribution to the extent
that the original distribution is skewed. But neither of these claims trans-
lates into the much wider claim that democracies will display either more
or less economic inequality than non-democracies. Considerations of this
sort help us to see that the apparently simple question ‘does democracy
reduce economic inequality?’ hides a range of further questions that in
turn require more detailed modelling. Such considerations will interest
normative political theorists too.
This article also reveals a further relationship between positive political
models and quantitative empirical work. As stressed previously, a key part
of the process of developing positive political models and theories
involves the ideas of abstraction, simplification and idealization. And
one key point here was the importance of being explicit about the abstrac-
tions, simplifications and idealizations involved in any model. When we
turn to quantitative empirical work, we need to specify not only the key
variables of interest (measures of economic inequality and measures of
democracy in this case), but also the control variables that might be
expected to influence economic inequality independently of the postu-
lated effects of the variable we are really interested in (democracy).
The link from the process of abstraction, simplification and idealization
to the specification of control variables should be clear. In a theoretical
model we can usefully abstract from an issue, but in a quantitative analysis
Positive Political Theory 213

we should control for the potential impact of that issue. So, being explicit
about the issues we abstract from in our theoretical work helps us to
identify the variables we will need to control for when we turn to
quantitative work.
By considering a range of different positive models, Timmons
exempliﬁes the idea of searching across models and considering variations
on a theme. But the mismatch between the rather narrowly focused
models considered and the broadly speciﬁed empirical investigation
might suggest that the theoretical models are rather ‘zoomed in’ on the
study of micro issues that arise within a democracy, while empirical work
has been rather ‘zoomed out’ on the macro questions that arise in com-
paring democracies with non-democracies. This in itself points to useful
directions for further work on both the theoretical and empirical fronts.

6 Conclusion
No ‘how-to’ guide to the practice of positive political theorizing can
guarantee that the theories/models generated will be useful, valuable
and interesting. But the steps outlined in the previous sections capture
and spell out the beneﬁts of thinking carefully and explicitly about the
positive theory/model element of any political analysis.
Early in this chapter I suggested that positive political theory had some
of the characteristics of prose, in that we use it all the time without always
being conscious of that fact. This is a metaphor that extends a little
further: just as the explicit study of grammar, syntax and punctuation
can improve our prose, whether as readers or writers, so the explicit study
of the more detailed structure of positive political theories and models
and the motivating and causal forces that they analyse can improve our
political debate and understanding, as both readers and writers. Just as
there are no perfect prose sentences, there are no perfect positive political
theories/models. But there are ways in which we can clarify meaning and
develop greater understanding, both in our prose and in our political
analysis.

References
Abrams, Samuel, Torben Iversen and David Soskice, 2011. ‘Informal social
networks and rational voting’, British Journal of Political Science 41, 229–57.
Aldrich, John, 1993. ‘Rational choice and turnout’, American Journal of Political
Science 37, 246–78.
Austen-Smith, David, and James Banks, 2000. Positive Political Theory I: Collective
Preference. Ann Arbor: University of Michigan Press.
214 Chapter 10

Austen-Smith, David and James Banks, 2005. Positive Political Theory II: Strategy
and Structure. Ann Arbor: University of Michigan Press.
Bendor, Jonathan, Daniel Diermeier and Michael Ting, 2003. ‘A behavioral model
of turnout’, American Political Science Review 97, 261–80.
Berggren, Niclas, Henrik Jordahl and Panu Poutvaara, 2009. ‘The looks of
a winner: beauty and electoral success’, Journal of Public Economics 94, 8–15.
Besley, Timothy and Stephen Coate, 1997. ‘An economic model of representative
democracy’, Quarterly Journal of Economics 112, 85–114.
Blais, Andre, 2000. To Vote or Not To Vote? The Merits and Limits of Rational Choice
Theory. University of Pittsburgh Press.
Brennan, Geoffrey and Alan Hamlin, 1998. ‘Expressive voting and electoral
equilibrium’, Public Choice 95, 149–75.
Brennan, Geoffrey and Hartmut Kliemt, 1994. ‘Finite lives and social institutions’,
Kyklos 47, 551–71.
Brennan, Geoffrey and Loren Lomasky, 1993. Democracy and Decision. Cambridge
University Press.
Buchanan, James, 1984. ‘Politics without romance: a sketch of positive public
choice theory and its normative implications’, in James Buchanan and
Robert Tollison, eds., The Theory of Public Choice II. Ann Arbor: University of
Michigan Press, 11–22.
Downs, Anthony, 1957. An Economic Theory of Democracy. New York:
Harper & Row.
Duverger, Maurice, 1965. Political Parties: Their Organization and Activity in the
Modern State. London: Wiley.
Elster, Jon, 1989. Nuts and Bolts for the Social Sciences. Cambridge University Press.
Ferejohn, John, 1986. ‘Incumbent performance and electoral control’, Public
Choice 50, 5–25.
Ferejohn, John and Morris Fiorina, 1974. ‘The paradox of not voting: a decision
theoretic analysis’, American Political Science Review 68, 525–36.
Fiorina, Morris, 1981. Retrospective Voting in American National Elections. New
Haven, CT: Yale University Press.
Fiorina, Morris, Samuel Abrams and Jeremy Pope, 2003. ‘The 2000
US presidential election: can retrospective voting be saved?’, British Journal of
Political Science 33, 163–87.
Forbes, Donald, 2004. ‘Positive political theory’, in Gerald Gaus and
Chandran Kukathas, eds., Handbook of Political Theory. London: Sage
Publications, 57–72.
Fowler, James, 2006. ‘Habitual voting and behavioral turnout’, Journal of Politics
68, 335–44.
Friedman, Jeffrey, ed., 1996. The Rational Choice Controversy: Economic Models of
Politics Reconsidered. New Haven, CT: Yale University Press.
Gordon, Scott, 1993. The History and Philosophy of Social Science. London:
Routledge.
Green, Donald and Ian Shapiro, 1994. Pathologies of Rational Choice Theory:
A Critique of Applications in Political Science. New Haven, CT: Yale University
Press.
Positive Political Theory 215

Hamlin, Alan and Colin Jennings, 2004. ‘Group formation and political conflict:
instrumental and expressive approaches’, Public Choice 118, 413–35.
Hamlin, Alan and Colin Jennings, 2007. ‘Leadership and conflict’, Journal of
Economic Behavior and Organization 64, 49–68.
Hamlin, Alan and Colin Jennings, 2011. ‘Expressive political behaviour: founda-
tions, scope and implications’, British Journal of Political Science 41, 645–70.
Hill, Lisa, 2010. ‘On the justifiability of compulsory voting: reply to Lever’, British
Journal of Political Science 40, 917–23.
Hindmoor, Andrew, 2006. Rational Choice. London: Palgrave.
Hindmoor, Andrew, 2011. ‘“Major combat operations have ended”? Arguing
about rational choice’, British Journal of Political Science 41, 191–210.
Kincaid, Harold, 1998. ‘Positivism in the social sciences’, in Edward Craig, ed.,
The Routledge Encyclopedia of Philosophy. London: Routledge, 558–61.
Krehbiel, Keith and John Wright, 1983. ‘The incumbency effect in congressional
elections: a test of two explanations’, American Journal of Political Science 27,
140–57.
Lever, Annabelle, 2010a. ‘Compulsory voting: a critical perspective’, British
Journal of Political Science 40, 897–915.
Lever, Annabelle, 2010b. ‘Democracy and voting: a response to Lisa Hill’, British
Journal of Political Science 40, 925–9.
Mueller, Dennis, 2003. Public Choice III. Cambridge University Press.
Müller, Wolfgang and Kaare Strøm, eds., 1999. Policy, Office, or Votes? How
Political Parties in Western Europe Make Hard Decisions. Cambridge University
Press.
Ordeshook, Peter, 1986. Game Theory and Political Theory: An Introduction.
Cambridge University Press.
Panebianco, Angelo, 1988. Political Parties: Organization and Power. Cambridge
University Press.
Riker, William and Peter Ordeshook, 1968. ‘A theory of the calculus of voting’,
American Political Science Review 62, 25–42.
Riker, William and Peter Ordeshook, 1973. An Introduction to Positive Political
Theory. Englewood Cliffs, NJ: Prentice Hall.
Satz, Debra and John Ferejohn, 1994. ‘Rational choice and social theory’,
The Journal of Philosophy 91, 71–87.
Shepsle, Kenneth and Mark Bonchek, 1997. Analyzing Politics: Rationality,
Behavior and Institutions. New York: Norton.
Strøm, Kaare, 1990. ‘A behavioral theory of competitive political parties’,
American Journal of Political Science 34, 565–98.
Tiebout, Charles, 1956. ‘A pure theory of local expenditures’, The Journal of
Political Economy 64, 416–24.
Timmons, Jeffrey, 2010. ‘Does democracy reduce economic inequality?’, British
Journal of Political Science 40, 741–57.
Van Evera, Stephen, 1997. Guide to Methods for Students of Political Science. Ithaca,
NY: Cornell University Press.
216 Chapter 10

Ware, Alan, 1992. ‘Activist-leader relations and the structure of political parties:
“exchange” models and vote-seeking behaviour in parties’, British Journal of
Political Science 22, 71–92.
Zuckerman, Alan, Nicholas Valentino and Ezra Zuckerman, 1994. ‘A structural
theory of vote choice: social and political networks and electoral ﬂows in Britain
and the United States’, The Journal of Politics 56, 1008–33.
11 Rational Choice Theory

Brian Kogelmann and Gerald Gaus

1 Introduction
‘The theory of justice is a part, perhaps the most significant part, of the
theory of rational choice’, writes John Rawls (1999a: 15). In linking the
theory of justice to rational choice Rawls both continued an intellectual
tradition and began a new one. He continued an intellectual tradition in
the sense that rational choice theory was first formally introduced in von
John von Neumann’s and Oscar Morgenstern’s Theory of Games and
Economic Behavior (1944), which launched a research agenda that con-
tinues to this day. Rawls, James M. Buchanan (1962; see Thrasher and
Gaus forthcoming), and John Harsanyi (1953, 1955) led the way in
applying the idea of rational choice to the derivation of moral and political
principles; a later generation of political theorists applied game theoretic
analysis to a wide variety of problems of social interactions (e.g. Hampton
1986; Sugden 1986; Binmore 2005). David Gauthier’s Morals by
Agreement (1986) was perhaps the most sustained and resolute attempt
to apply models of utility maximization and decision theory to the deriva-
tion of social morality. Yet, despite the fact that some of the most respected
political theorists of the past fifty years extensively employed the tools of
rational choice theory, confusion still reigns about what these tools pre-
suppose and how they are to be applied. Many reject the entire approach by
a common refrain that ‘rational choice assumes that people are selfish’, and
since that is false, the entire line of analysis can be dismissed; a slightly
more sophisticated (but still misplaced) dismissal insists that rational
choice is about ‘preferences’, but political philosophers are interested in
‘reasons’, so again the entire approach can be set aside asserting that
rational choice theory assumes selfishness, denying that people are selfish,
and dismissing rational choice accordingly. A fundamental aim of this
chapter is to explain just what the tools of rational choice theory are –
and what presuppositions they make – and to provide some guidance for
those wishing to follow in Rawls’s footsteps, showing how rational choice

217
218 Chapter 11

theory can be applied to some of the fundamental issues of moral and

political theory.
But even for those who do not want to explicitly engage in the
contractarian project and thus follow in Rawls’s footsteps, there are still
important reasons to learn the basics of rational choice theory. First, one
cannot properly engage with those who do explicitly rely on rational
choice theory without understanding its basics. As just one example,
arguing that parties in Rawls’s original position would not choose the
difference principle is fruitless unless one further engages with how Rawls
defines the original position – given the tools of rational choice theory
Rawls relies on, and given how the choice problem is defined, the only
rational choice is the difference principle. Second, even if one does not
want to primarily use rational choice theory in one’s theorizing,
knowledge of the theory can open up avenues of research to lend
a supporting role. As an example, even though a deliberative democrat’s
core normative theory does not rely on rational choice in the way that
Rawls’s does, knowledge of the theory can help one engage, say, in the
social scientific institutional design literature that could be relevant for
the democrat’s project.
In Section 2 we explain preferences and utility functions as well as
parametric and strategic choice. Section 3 sketches the dispute about
whether rational choice is an adequate mode of explanation, while
Section 4 highlights the way in which rational choice theory is a normative
theory, and how it has been employed in normative political philosophy.
Section 5 seeks to draw our observations together in a checklist of
decisions for those who would employ rational choice methods in their
own work.

2 What Is Rational Choice Theory?

2.1 Preferences and Utility Functions

Rational choice is, at bottom, a theory of preference maximization. Many
think of preferences as non-relational tastes or desires – having
a preference for something is simply liking or desiring that thing.
On this common view, to say that Alf has a preference for mangos
means that Alf has a taste or desire for mangos; thus to say that he
maximizes his preferences is just to say that he maximizes his pleasures
or desire satisfaction. It is thus very common – for both defenders and
critics – to see rational choice theory as inherently bound up with
a Hobbesian or Benthamite psychology (Plamenatz 1973: 20–7,
149–59; Kliemt 2009: 46ff. Sen 2009: 178–83). While such an
Rational Choice Theory 219

Table: 11.1

μ function A μ function B μ function C

Mangos 3 10 1,000
Bananas 2 5 99
Apples 1 0 1

understanding of preferences is perhaps colloquially natural, rational

choice theory always understands preferences as comparative or rela-
tional, and they have no necessary connection to desires or pleasure.
The core of rational choice theory is the primitive conception of
a preference as a binary relation, of the form ‘x is preferred to y’
(x⊱y). This is emphatically not a comparison of the strengths of two
preferences, that for x and for y. It is literally incoherent in decision
theory to claim one simply ‘prefers mangos’ – a preference is a binary
comparison.
Perhaps the best understanding of preferences is as deliberative
rankings over states of affairs. When Alf deliberatively ranks (x), the
state of affairs in which he eats a mango over (y), the state of affairs in
which he eats a banana, we say that Alf prefers mangos to bananas (x⊱y).
Note that we can leave entirely open the considerations Alf employed to
arrive at this ranking: these could have been self-interest, desire satisfac-
tion, or his conception of virtue (‘The brave eat mangos, while only
cowards eat bananas!’). Utility theory is a way to represent consistent choice,
not the foundation of those choices. If Alf employs some deliberative
criteria such that he ranks x above y for the purposes of choice, then we
can represent this as x⊱y; it is critical to realize that he does not rank
x above y because he prefers x to y.
These preferences over states of affairs, combined with further infor-
mation, can be used to generate preferences over actions (say, α and β).
Thus in rational choice theory we can map preferences over ‘outcomes’
(x, y) to preferences over actions (α, β), which, at least as a ﬁrst approx-
imation, can be understood as simply routes to states of affairs. Does Alf
prefer grocery shopping at (α) Joseph’s or (β) Caputo’s? That depends on
his information. If Alf knows that Joseph’s has only mangos (x) and
Caputo’s has only bananas (y), then since x⊱y, α⊱β – he will prefer
shopping at Joseph’s to shopping at Caputo’s.
From an individual’s preferences we can derive an ordinal utility
function. Let us consider preferences over three fruits: mangos (x),
bananas (y), and apples (z). An ordinal utility function is a numerical
220 Chapter 11

representation of a person’s preferences, examples of which are illustrated

in Table 11.1. We can derive an ordinal utility function for any individual
so long as the individual’s preferences satisfy the following axioms.
1. Preferences are complete. For any two states of affairs x and y, either
x⊱y, y⊱x or, we can say, Alf is indifferent between x and y (x∼y).
We can define indifference in terms of what might be called the true
primitive preference relation, ‘at least as good as’. If Alf is indifferent
between x and y (x∼y), we can say that x is at least as good as y (x≿y)
and y is at least as good as x (y≿x). We can also define ‘strict prefer-
ence’ (x⊱y) in terms of this more fundamental binary relation ‘at least
as good as’: x⊱y implies x≿y and ¬(y≿x).1
2. Strict preferences are asymmetric. x⊱y implies ¬(y⊱x). Indifference is,
however, symmetric: (x∼y) implies (y∼x).
3. The true primitive preference is reflexive. Alf must hold that state of
affairs x is at least as good as itself (x≿x).
4. Preferences must be transitive. If Alf prefers x to y and y to z, then Alf
must prefer x to z (x≻y & y≻z implies x≻z).
Note that the numbers used to rank options in Alf’s ordinal utility
function tell us very little. All an ordinal utility function implies is that
higher-numbered states of affairs are more preferred than lower-
numbered states of affairs. Turning back to Table 11.1, utility function
A contains the same information as utility function B which contains the
same information as utility function C. They all tell us that Alf prefers
mangos to bananas to apples, nothing more. Thus ordinal utility
information can be limiting. As we shall see, when modeling rational
choice under risk and uncertainty, we require information about (to put
the matter rather roughly) the ‘distances between preferences’. We need
to know more than the fact that Alf prefers mangos to bananas – we must
know (again, very roughly) how much more he prefers mangos to bana-
nas. This brings us to cardinal utility functions, which contain such infor-
mation. We can take any ordinal utility function and derive a cardinal
utility function so long as preferences satisfy a few more axioms:2
5. Continuity. Assume again that for Alf mangos (x) are preferred to
bananas (y), and bananas are preferred to apples (z), so x≻y, y≻z.
There must exist some lottery where Alf has a p chance of winning his
most preferred option (x) and a 1-p chance of receiving his least
preferred option (z), such that Alf is indifferent between playing that

1
Read ‘¬’ as ‘not’.
2
For simplicity’s sake, we state these in terms of the strict preference relation. A more
general statement would use the ‘at least as good’ relation. The intuitive ideas are clearer
with strict preference.
Rational Choice Theory 221

lottery and receiving his middle option (y) for sure. So if such a lottery
is L(x,z), we can say that for Alf [L(x,z)]∼y.
6. Better Prizes. When Alf is confronted with two lotteries, L1 and L2,
which (i) have the same probabilities over outcomes (for example, 0.8
chance of the prize in the first position, and so 0.2 chance of the prize
in the second position), (ii) the second position has the same prize in
both L1 and L2, and (iii) in the first position, L1’s prize is preferred by
Alf to L2’s, Alf must prefer playing lottery L1 to lottery L2 (L1≻L2).
7. Better Chances. When Alf is confronted with two lotteries, L1 and L2,
where (i) L1 and L2 have the same prizes in both positions (x and y,
where x≻y) and (ii) L1 has a higher chance (say 0.8) of winning x than
L2 (say 0.6), Alf must prefer playing lottery L1 to lottery L2 (L1≻L2).
8. Reduction of Compound Lotteries. Alf’s preferences over compound
lotteries (where the prize of a lottery is another lottery) must be
reducible to a simple lottery between prizes. Thus the value of winning
a lottery as a prize can be entirely reduced to the chances of receiving
the prizes it offers – there is no additional preference simply for
winning lottery tickets (say, the thrill of lotteries).
If Alf’s preferences obey these axioms, we can derive a cardinal utility
function in the following manner, originally proposed by von Neumann
and Morgenstern (1944: ch. 3). Consider Alf’s ordinal utility function,
represented in Table 11.1. We take Alf’s most preferred option, x, and
assign it an arbitrary value, say 1. We then take Alf’s least preferred
option, z, and assign it an arbitrary value, say 0. We then take an option
in between (y) and compare Alf’s preference between that option and
a lottery between x and z. We manipulate the probability of the lottery
until Alf is indifferent between that lottery and the middling option under
consideration (the Continuity axiom guarantees there will be such lot-
tery). Suppose Alf is indifferent between y and a lottery with p = 0.6
chance of winning x and, so, a 1-p (0.4) chance of z (the probabilities in
the lottery must sum to 1). Because indifference obtains at this specific
probability, we assign a numerical value of 0.6 to the state of affairs in
which Alf receives y, a banana. Alf’s new cardinal utility function, repre-
senting how much more he prefers certain outcomes to others, is shown in
Table 11.2.
Just how we are to understand this information is disputed. On a very
strict interpretation, all we have found out is something about Alf’s
propensity to engage in certain sorts of risks; on a somewhat – but not
terribly – looser interpretation we have found out something about the
ratios of the differences between the utility of x, y, and z, which can be
further interpreted as information about the relative intensity of Alf’s
three preferences. It is important that our (1, 0.6, 0) utility scale is simply
222 Chapter 11

Table: 11.2

μ function

Mangos 1
Bananas 0.6
Apples 0

a representation as to how Alf views the relative choice worthiness of the

options; he does not ‘seek’ utility, much less to maximize it. He seeks
mangos, bananas, and apples. To understand his actions as maximizing
utility is to say that his consistent choices can be numerically represented
so that they maximize a function. It is also critical to realize that the only
information we have obtained is a representation of the ratios of the
differences between the three options. There are an inﬁnite number of
utility functions that represent this information. Cardinal utility functions
are thus only unique up to a linear transformation (which preserves this
ratio information). If we call the utility function we have derived U, then
any alternative utility function U', where U' = aU +b (where a is a positive
real number and b is any real number), contains the same information.
We can readily see why these utility functions are not interpersonally
comparable: it makes little sense to simply add the utility functions of Alf
and Betty, when each of their functions is equally well described in an
inﬁnite number of ways, with very different numbers.

2.2 Parametric Choice

After characterizing an agent’s utility function we can go on to examine
what sorts of choices rational agents make. The basic idea behind rational
choice theory is that choosers maximize expected utility. Since utility, as we
have seen, is simply a numerical representation of an agent’s preferences,
maximizing expected utility means that a rational agent maximizes the
satisfaction of her preferences. Although this sounds straightforward –
and it is, in certain contexts – maximizing one’s preferences can be quite
complex once we start examining choice under risk and uncertainty.
Let us first consider rational parametric choice. When an agent chooses
parametrically, it is assumed that the choices that the agent makes do not
influence the parameters of other actors’ choices (i.e., do not influence
their preferences over outcomes) and vice versa. Their combined choices
can affect her options (for example, the combined choices of other con-
sumers determines prices), but when she acts, she takes all this as fixed
Rational Choice Theory 223

and beyond her control. As far as the chooser is concerned, she has a ﬁxed
preference ordering and confronts a set of outcomes mapped onto a set of
actions correlated with those outcomes. This is opposed to strategic
choice, examined in the next section. We can understand decision theory
as the study of rational parametric choice, game theory the study of
rational strategic choice.
In the simplest case of rational choice, the individual is choosing
under certainty. Here Betty not only knows her orderings of outcomes
(x⊱y⊱z) – an assumption we make throughout – but also that she knows
with certainty that, say, action α produces x, action β produces y, and
action γ produces z. Given this, she has no problem ordering her action
options α⊱β⊱γ, and so as a rational utility maximizer she chooses α out of
the set of options (α, β, γ). But we do not always choose under certainty.
Sometimes we choose under risk. When we are choosing under risk, the
outcomes of our action options are not certain, but we do know the
probability of different outcomes resulting from our choice act. Suppose
Alf faces the option between action α (choosing covered fruit bowl A)
and β (covered fruit bowl B). Bowl A might contain a mango or it might
contain an apple – Alf does not know which. Bowl B might contain
a mango or it might contain a banana. Once again Alf does not know
which. Alf does know, however, that if bowl A is chosen, then there is
a 0.5 chance of getting a mango and a 0.5 chance of getting an apple. Alf
also knows that if bowl B is chosen, then there is a 0.25 chance of getting
a mango and a 0.75 chance of getting a banana. Because Alf knows the
probabilities that each action will produce different outcomes, Alf is
choosing under risk. How does a rational person choose in risky
situations?
In such situations, rational agents use an expected utility calculus. With
an expected utility calculus, agents multiply the probability an action that
will produce an outcome by the utility one assigns to that outcome.
We assume that Alf knows all the possible outcomes that each action
will produce. Suppose Alf has a von Neumann and Morgenstern cardinal
function over outcomes as depicted in Table 11.2. With this cardinal
utility function, Alf’s expected utility calculus goes like this:
α: choosing bowl A: 0.5(1) + 0.5(0) = 0.5
β: choosing bowl B: 0.25(1) + 0.75(0.6) = 0.7
Since the expected utility of α (choosing bowl A) is 0.5 and since the
expected utility of β (choosing bowl B) is 0.7, Alf chooses what is the best
prospect for maximizing his preferences, which is β. Notice that this
does not ensure that, in the end, he will have achieved the highest level of
preference satisfaction: if he did choose α and it turned out to have the
mango while β produced the banana, α would have given him the best
224 Chapter 11

outcome. The point of expected utility is that, at the time of choice, β

offers the best prospects for satisfying his preferences.
It is sometimes thought that any set of cardinal numbers is sufﬁcient
for expected utility calculations, as we can multiply cardinal numbers by
probabilities (you can’t multiply ordinal utilities). This is too simple.
Our von Neumann–Morgenstern cardinal utility functions possess the
expected utility property, as they take account of Alf’s attitude toward risk.
Recall that von Neumann–Morgenstern functions are generated
through preferences over lotteries, thus they include information
about how Alf weighs the attractiveness of an outcome and the risk of
failing to achieve it. Suppose that instead of employing the von
Neumann–Morgenstern procedure, we asked Alf to rate, on a scale of
0 to 1, simply how much he liked mangos, bananas, and apples, and he
reports 1, 0.6, and 0 (this looks just like Table 11.2). If we then multi-
plied, say, this type of ‘cardinal utility’ of 1 for mangos times a 0.5
probability that α will produce a mango, we would get the expected
value of α, but not its expected utility, for Alf could be very reluctant
to take a 0.5 chance of his worst outcome (an apple) in order to get a 0.5
chance of his best; so to him the utility of α could be far less than 0.5. But
our von Neumann–Morgenstern procedure already factored in this
information. Only in the special case in which people are risk-neutral
(they are neither risk prone nor risk averse) will expected value be
equivalent to expected utility.
Choice under uncertainty is a messier and more complex affair. Like
choice under risk, when choosing under uncertainty, one does not know
the outcome of one’s act of choice. But unlike choice under risk, one also
does not know the probabilities that an action will yield the various
possible outcomes. Following R. Duncan Luce and Howard Raiffa
(1957: 278), we note that the idea of uncertainty is vague. With choice
under risk, we know with objective certainty the relevant probabilities that
α will produce various outcomes (0.5 of x, 0.5 of z). What happens,
though, if we are not certain about the relevant probabilities but more
or less have a reasonable guess as to what the probabilities are? Is this
choice under uncertainty? Technically speaking, it is. Instead of having
two categories, risk and uncertainty, it is more helpful to think of uncer-
tainty as coming in degrees. Though we do not know the probabilities of
outcomes occurring, our credence in what the relevant probabilities are
can be more or less founded. At one end of the spectrum is choice under
risk, where we know the probabilities of any given action option (so that
this sums to 1 for each act), and on the other end of the spectrum is choice
under radical uncertainty, where we cannot even begin to intelligently
guess the relevant probabilities over outcomes.
Rational Choice Theory 225

Most relevant for political theory is choice under radical uncertainty.

As we shall see later on, one of the central debates in twentieth-century
political theory largely revolved around what rational choice requires
when one faces such uncertainty. Suppose, as before, that Alf must
choose between α and β (on the basis of their associated fruit bowls).
But unlike before, Alf has absolutely no basis for assigning probabilities as
to which fruit the respective bowls might contain. In such a situation how
does Alf choose rationally? It is here that rational choice theory turns more
into an art than a deductive system. Although it is relatively straightfor-
ward how rational persons choose under risk, and somewhat less certain
under uncertainty, it is by no means obvious how rational agents should
choose when faced with radical uncertainty. Consider two different ways
in which one might approach such choices, outlining the costs associated
with the different methods.
(i) When facing radical uncertainty one might simply try to extend
the choice procedure used in decision under risk – that is, one might
employ an expected utility calculus. But an expected utility calculus
requires that we assign probabilities to outcomes and, by hypothesis,
when facing radical uncertainty, there is no educated basis for
assigning such probabilities. In such cases the Laplacean principle of
indifference says to assign equal probability to all possible outcomes
occurring, reﬂecting indifference to these outcomes occurring.
Continuing our example, using this Laplacean principle, Alf would
assign an equal probability of α yielding a mango and an apple, and to
β an equal probability of it giving him a mango and a banana. From
there Alf performs his expected utility calculus as he did with choice
under risk, so α = 0.5(1) + 0.5(0); β= 0.5(1) + 0.5(0.6), again yielding
the choice of β.
The Laplacean principle of indifference is controversial. Ken Binmore
(2009: 129) points out the ambiguity present in assigning equal
probability to different outcomes occurring. Suppose Alf attends a three-
horse race between α, β, and γ, but has no basis for assigning probability as
to which horse will win. Is the Laplacean principle of indifference to be
applied to the question (a) ‘which horse will win?’ or (b) ‘will α win or not
win?’, ‘will β win or not win?’, ‘will γ win or not win?’. If the ﬁrst, the
Laplacean principle says Alf should assign each horse a ⅓ probability of
winning (so that Alf is indifferent between which horse wins).
If the second, the principle says to assign each horse a ½ probability of
winning and ½ probability of not winning (so that Alf is indifferent
between each horse winning and not winning).
(ii) Another way of choosing under radical uncertainty is by using the
maximin decision rule. The maximin decision rule eschews the use of
226 Chapter 11

Prize 1 Prize 2
L1 0 n

L2 1/n 1

Figure 11.1: Rawls’s problem for maximin

probabilities; it instructs one to examine each action option (α, β, γ) and

identify for each its worst possible outcome (say αW, βW, and γW).
Looking only at these ‘minimum payoffs’ for each action one is to select
the highest among them (hence one should ‘maximize the minimum’).
Like the Laplacean principle of indifference, there are many criticisms of
the maximin decision rule, most of which offer examples in which its
application seems irrational. Suppose, following Rawls (1999: 136), one
is faced with a choice between lottery L1 and lottery L2 whose probabil-
ities are unknown. Suppose further the payoffs for L1 and L2 are in dollars
(and that preferences are monotonic over money). This decision problem
is shown in Figure 11.1.
Maximin says to choose L2 because L2 guarantees a minimum payoff of
1/n while the minimum payoff for L1 is 0. And when n is small (say, $1,
$2, or $3), this does not seem to be an unreasonable choice. But suppose
n is quite large, perhaps $1,000,000. In this case, there does seem some-
thing irrational about choosing L2 and thus something irrational about
the maximin decision rule in general.

2.3 Strategic Choice

When choosing parametrically, the decisions of others are a fixed back-
ground; one does not have to include in one’s deliberations the different
things they might do. In such cases of parametric choice, maximizing
one’s preferences is relatively straightforward, at least until we get to
choice under uncertainty. But many times the actions of agents affect
each other’s decisions. This is strategic choice, which is studied by game
theory. Roughly, a game is defined by (i) a set of players, (ii) a set of
strategies for each player, (iii) the information available to the players, and
(iv) a set of payoffs for each player which is defined by her preference
ordering (ordinal or cardinal) which are mapped onto her strategies.
Change any of these and you change the game.
Game theorists originally focused on zero-sum games, interactions of
pure conflict. Whatever one player wins the other player loses. To see this,
consider Figure 11.2, which is the classic game Matching Pennies.
Rational Choice Theory 227

Betty
Heads Tails
α β
Heads 1 –1
α –1 1
Alf Tails –1 1
β 1 –1

Figure 11.2: Matching Pennies

By convention, Row’s (Alf’s) payoffs are in the bottom left of each cell,
Column’s (Betty’s) in the upper right. Note that the gains and losses in every
cell equal 0; any gain must come from the other player, so we have a zero-
sum game. John von Neumann (1928) proved that all zero-sum games with
finite strategies have equilibrium solutions. That is, there is a rational course
of action in all such games. Relating back to decision under uncertainty in
parametric choice, the equilibrium solution for zero-sum games is to play
one’s maximin strategy – to maximize one’s minimum. If both players
maximize their minimums, then neither player can gain from unilaterally
changing her strategy. If we accept that the equilibrium solution concept
models rationality, then, when playing zero-sum games, it is rational to
maximin.
Though the theory of games was first developed in zero-sum form, it
was quickly realized that zero-sum games are of limited applicability for
understanding most strategic interactions. Rarely are we locked into total
conflict; we often find ourselves in situations where we are competing and
cooperating. Consider, for example, Luce and Raiffa’s (1957: 90–1)
politically incorrect ‘Battle of the Sexes’ game. In this game, Alf and
Betty wish to go out with each other on a Friday evening; the worst
thing for each would be to not go out together. However, Alf would prefer
a date at the prizefights, Betty at the ballet. Figure 11.3 gives an example
of the game, using cardinal utility.
So we have two players, Alf and Betty, and each has the same two ‘pure’
strategies, {Alf α, Alf β}, {Betty α, Betty β}. We suppose that each knows
the information set S: they know each other’s utilities and what strategies
are open to the other, each has knowledge of the other’s rationality, and
they know that each must choose without knowing what the other has
chosen. Moreover they each know S': each knows that the other knows all
the facts in S. And indeed they know S": each knows that the others know
all the facts in S'. The common knowledge assumption is assumed to be
iterative – we can go on with higher and higher levels, where each knows
what the facts were at the previous level.
228 Chapter 11

Betty
Go to Go to
Fights Ballet
α β
Go to 1 0
Alf Fights 2 0
α
Go to 0 2
Ballet 0 1
β

Figure 11.3: The ‘Battle of the Sexes’ in cardinal utility

If Alf goes to the fights, then Betty, given her own preferences, will do
best by also going to the fights; if Betty goes to the ballet, Alf will do best
by also going to the ballet. Neither knows what the other is doing, but
they wish to coordinate. So this is a game of cooperation. Yet it is also
a game of conflict: they disagree about the best outcome. The central
concept in game theory is that of Nash equilibrium. We can think of the
Nash equilibrium as the solution to a game insofar as when two players
are playing their equilibrium strategies each has made his/her best
response to the move of the other player, and so neither player has any
incentive to unilaterally change his/her strategy. The strategic
interaction, we might say, can come to rest there. In the ‘Battle of the
Sexes’ game, there are two Nash equilibria in pure strategies: both play α
or both play β. Players can also play a ‘mixed strategy’, which is
a probability distribution over their pure strategies. In Figure 11.2’s
game, this would mean Alf would play α with a p probability and β
with a 1-p probability. This game has a Nash equilibrium in mixed
strategies in which Alf plays α with a probability of ⅔ (and so β with
a probability of ⅓) and Betty plays β with a probability of ⅔. At this
point, no adjustment to the probability that they play each of their pure
strategies will allow a player to unilaterally improve her expected
payoffs. John Nash (1950a, 1951), in a fundamental theorem of game
theory, showed that all games in cardinal utility with finite strategies
have at least one equilibrium (sometimes only a mixed equilibrium).
Note that this theorem, because it requires mixed strategies, depends on
the expected utility property.
We should note a more general solution concept, correlated equilibrium
(Aumann 1987), which includes the Nash solution. Its attractions are
manifest in the game of Chicken:
Rational Choice Theory 229

Chevy
Swerve Don’t
α Swerve
β
Swerve 5 10
Ford α 5 0
Don’t 0 –5
Swerve 10 –5
β

Figure 11.4: Chicken in cardinal utility

Suppose that every Saturday and Sunday, two teenagers, one driving
a Chevy and the other a Ford, drive toward each other with the pedal to
the metal, and the first one to swerve is ‘chicken’. So the options are
between swerving (α) and not swerving (β). The best result is to keep
driving straight while the other swerves: the other chickened; you didn’t.
If both swerve, neither crashes; if neither swerves, they crash. In this
game, the two ‘pure equilibria’ are (Ford α, Chevy β) and (Ford β;
Chevy α). There is a mixed Nash equilibrium where each swerves half
the time, but this has the unfortunate result that ¼ of the time they crash!
Now consider a game where we add a strategy: on Saturdays Chevy
swerves, on Sundays Ford swerves. The game will not be terribly exciting,
but they will never crash, and each will have an expected payoff of 5, as
good as one could do against a rational opponent. But note that we
seemed to have changed the game: we have expanded the strategy set to
include a correlating device (the day and the make of car).
At least in their simpler forms, Chicken and the ‘Battle of the Sexes’
raise problems when we understand strategic rational choice as requiring
agents to play their equilibrium strategies. For in these games there are
multiple equilibria. What, then, do rational players do? Which equilibrium
do they play? Unfortunately there is no definitive good answer. People
have tried many ways to select one equilibrium from multiple Nash
equilibria. The oldest way of solving the equilibrium selection problem
is to a priori theorize about what rationality requires when faced with
multiple equilibria. This is known as ‘equilibrium refinement’ and was
pursued to its fullest by John Harsanyi and Reinhard Selten (1988). Some
refinements seem plausible: for instance, if one equilibrium strictly dom-
inates another (i.e., is better for all parties), then perhaps that equilibrium
is more rational. At best, this is a rather limited refinement. As we
230 Chapter 11

proceed, we encounter deep controversy – no satisfactory account has

been given to refine all games down to a unique equilibrium.
Others have employed evolutionary accounts explaining why some
equilibria are selected over others: though a pair of strategies might be
in equilibrium for a one-shot game, such a pair of strategies might no
longer be in equilibrium if the game is continually repeated (the opposite
is also true). Notable here is Brian Skyrms’s (1996, 2004) work.
In contrast, Thomas Schelling (1960: 57) emphasized the role of salience
or the sheer obviousness of a solution in certain contexts. For example,
when two parties are asked to split up a sum of money, they will often split
it right down the middle, even though there are many possible splits. Or if
two parties are asked to meet at some place in Paris, they will usually go to
the Eiffel Tower because it is such a large, obvious landmark. Of course,
there can be multiple salient equilibria. Moreover, differing methods of
equilibrium selection can conflict. Strict dominance as a plausible
refinement technique can conflict with Schelling’s salience – sometimes
the salient equilibrium will be dominated by a non-salient equilibrium.
Though there have been many proposals of how to select among multiple
equilibria, a current limitation in the theory of games is that it cannot tell
us which equilibrium rational players select.
One final area of strategic choice relevant for political theory is bargain-
ing games. Implicitly we have been examining non-cooperative game
theory, which assumes that choices among strategies are not binding.
One follows a Nash equilibrium not because it is the solution to a game,
but because, given the move of the other, it is one’s best available move.
In cooperative game theory, choices among strategies are assumed to be
binding by some external enforcement mechanism. The most famous
case of a cooperative solution that is not in equilibrium is the Prisoner’s
Dilemma. The story is normally told in terms of prisoners getting caught
and seeking to minimize jail time, but all games are really about utility.
Figure 11.5 gives the Prisoner’s Dilemma in terms of cardinal utility.
If Betty plays α, Alf will get x if he plays α, and 1 if he plays β; since 1 > x,
he should then play β. If Betty plays β, Alf gets 0 if he plays α, and y if he
plays β; since y > 0, he should again play β. ‘Confessing’ (β) is thus his
dominant strategy: no matter what Betty does, Alf does best if he plays β.
And it can clearly be seen that Betty will reason in exactly the same way.
So (β,β) is the sole Nash equilibrium. Yet the outcome of y/y is worse for
each than the x/x outcome. The x/x is a cooperative outcome that is not in
equilibrium; as a non-cooperative game there is only one solution to the
Prisoner’s Dilemma: confess! Such games can, however, form the basis of
binding agreements that yield a cooperative surplus. Faced with such
bargains, one might ask what distribution of surplus rational players will
Rational Choice Theory 231

Betty
Don’t Confess
Confess β
α
Don’t x 1
Confess x 0
Alf α
Confess 0 y
β 1 y
Where 1>x>y>0

Figure 11.5: The general form of the Prisoner’s Dilemma in cardinal

utility

settle on. The status of bargaining theory is unsettled. There are several
competing bargaining theories. Nash’s (1950b) bargaining solution has
the widest support. And Ariel Rubenstein (1982) appeared to put
cooperative bargaining theory on a ﬁrmer footing by clearly showing
how it can be derived from non-cooperative game theory. Yet Luce and
Raiffa’s (1957: 119–20) worry remains – that such bargaining solutions,
though intellectually interesting, do not seem to model what purely
rational players would select in actual bargains. And some worry that
the very axioms of bargaining theory go beyond rationality to include
some criteria of fairness (Thrasher 2014).

3 Is Rational Choice Theory Unrealistic?

3.1 People Aren’t Always Rational

It should be manifest that rational choice theory idealizes: agents are
generally understood to know all their options, rank them, and usually
in a way that provides a cardinal scale. They are then assumed to be
thoroughly consistent choosers, always seeking to maximize the satisfac-
tion of their preferences. The most common refrain is that, since people
are not thoroughly rational, these are wildly unrealistic assumptions, and
so undermine the usefulness of rational choice theory. If people aren’t
rational, why care about rational choice?
One version of this common refrain simply points out that people are
not always rational: they do not always act as rational choice theory
predicts. First, we need to be clear about what rational choice theory
does not predict: it does not predict that people will always be self-
interested, or that they seek to maximize the satisfaction of their desires,
232 Chapter 11

or that moral considerations never move them. These are matters about
the nature of the deliberation that determine a person’s rankings – they
concern the basis on which Alf determines whether state of affairs x is to
be ranked above y. As we have stressed, rational choice theory is not
committed to any account about the basis of rankings. Recall also that
preferences are not desires, but rankings of outcomes, which determine
preferences over actions. This is important. Many experimenters, for
example, believe that they have shown that people cooperate in
Prisoner’s Dilemmas. In experiments in which people are confronted
with situations that mimic Figure 11.5, where the payoffs are monetary
and players are awarded sums of money where the size of the monetary
payoffs is exactly correlated with the utility payoff schedule in
Figure 11.5, we can observe people cooperating – an action that, for all
players, is dominated by confessing. However, unless the players rank the
outcomes only in terms of their monetary payoffs – unless their prefer-
ences are only over amounts of money, and thus, for example, have no
concern with fairness or simply not being a jerk – we cannot say that the
players are actually in a Prisoner’s Dilemma interaction, so rational
choice theory does not predict that they will defect.
Nevertheless, even when we are more careful, there is dispute over
whether rational choice theory is a good predictor. Perhaps the most
comprehensive critique of rational choice theory as a predictive tool is
Ian Shapiro and Donald Green’s (1996) Pathologies of Rational Choice
Theory (see also Amadae 2015), though there have been equally
comprehensive defenses of rational choice theory as a predictive tool
(see, for instance, the debate in Friedman 1996). Rational choice theory
is a model of human action: like all models, we build a simpliﬁed world so
that we can better understand some salient dynamics, and get some
predictive leverage in some contexts. As Michael Wesiberg (2007)
notes: ‘[w]hen faced with . . . complexity, theorists can employ one of
several strategies: They can try to include as much complexity as possible
in their theoretical representations. They can make strategic decisions
about which aspects of a phenomenon can be legitimately excluded from
a representation. Or they can model, studying a complex phenomenon in
the real world by ﬁrst constructing and then studying a model of the
phenomenon.’ These models abstract from the complexity of reality to
analyze the interactions among key elements. In some contexts, the
models may abstract from too much of the complexity to give adequate
understanding of the phenomena, but in other contexts they can be
enlightening and accurate. In social science, no model or theory explains
most of the variance of interestingly complex phenomena, but this does
not mean the model is without predictive and explanatory power.
Rational Choice Theory 233

Rational choice models do rather well in a variety of cases, especially

where we can make accurate suppositions as to what people’s utility
functions look like, the information sets they have, and the actions open
to them. Thus experiments have conﬁrmed that various sorts of auction
behavior and trading behavior conform to the predictions of rational
choice theory (Plott 1986; Smith 2000). On the other hand, when people
have complex and heterogeneous utility functions, when the options are
not well understood, or, more radically, when they act out of strong
emotions, rational choice theory is not apt to explain things well
(Schwartz 2002; Bosman, Sutter, and Van Winden 2005).

3.2 Rationality and Intelligibility

A more radical rejection of rational choice theory insists that rationality is
a ‘folk concept’, not a scientiﬁc one. While most of us, as normal ‘folk’,
think in terms of preferences, options, and choices, some argue a ‘true’
scientiﬁc theory would be purely causal, and so would explain human
actions in terms of psychological or biochemical processes. Hartmut
Kleimt advances what might be understood as a ‘compatibilist’ reply:
‘humans are citizens of two worlds, a lower “material” world and a higher
one of reason’ (2009: 16). On this compatibilist approach, we can under-
stand humans both in simply causal terms and in purposive ways – as
agents with preferences – who seek to secure their most favored outcome.
Some suggest that we only rely on rational choice models because our
causal models are underdeveloped, so perhaps someday we can do with-
out rational choice theory. But this seems wrong. Humans understand
each other through ‘mind reading’: the actions of another are not really
intelligible to me until I see how, if I thought what she believes and I cared
for what she cared for, I would have done what she did. Making others
intelligible to us is closely bound to seeing them as rational: rational action
makes sense to us (Rosenberg 1995). True, sometimes it is intelligible to
us why people are not rational: we can understand all too well why
someone who is drunk accepts a dangerous and silly dare. But usually,
when we are confronted by simply irrational behavior, we do not
understand what it is really all about. To understand the behavior of
another is to see it as intelligible, and to see it is rational is to render it
intelligible.

4 Rational Choice as a Normative Theory

Herbert Gintis (2009: chs. 1, 12) boldly conjectures that game theory’s
axioms of consistent choice express, at the deepest level, the ingredients of
234 Chapter 11

evolutionary success. It is indeed striking how powerful, in the guise of

‘evolutionary game theory’, the tools of game theory have proven in
modeling natural selection (for classic statements, see Hamilton 1964;
Trivers 1971; Smith 1975). If the basic idea of consistent choosers who
maximize their satisfaction of their goals is foundational to selected
behavior in humans and non-humans alike, we might hypothesize that
creatures like us – who are conscious of their agency – cannot help but see
these axioms as normative: as not simply explaining what they do, but
guiding what they should do. In a secular world where the gods no longer
seem to speak to and guide us, can we have any deeper guide than our
most basic axioms of successful agency, on which perhaps our entire
evolutionary past hinges? Thus is the crux of Gauthier’s (1991) case for
what he calls ‘contractarianism’: the only truly solid foundation for think-
ing about what we ought to do that our naturalistic and scientiﬁc age has
left us are the consistency axioms at the root of successful agency. If we
have any hope of identifying normative principles for our post-
metaphysical world – a world shorn of faith in God’s plan for us, or
invisible normative properties that instruct us what to do – they must be
solidly grounded in rational choice.
Contractarianism in moral and political theory seeks to show that
moral and political principles or rules can be the result of an agreement
between rational parties and, further, that compliance with these princi-
ples or rules is rational. When deriving normative principles, contractar-
ians appeal to what rational parties would agree to, which is just to say
what rational choice requires in strategic interaction. This can be seen
informally in the case of Hobbes when he argues that rational parties,
engaged in a war of all against all in the state of nature, would agree to
enter the social contract, where each gives up her right to all things in
exchange for security from the sovereign (1994: chs. 13, 16).
Contemporary contractarians go further than examining the informal
logic of what rational parties would agree to. Gauthier (1986), for
instance, employs game theoretic bargaining models to examine what
moral principles rational parties would agree on. He ends up developing
his own bargaining solution, the principle of minimax relative concession,
which is an offshoot of the Kalai-Smorodinsky (1975) bargaining solu-
tion. Most contractarians – including Gauthier at one point in his career –
opt for the alternative, Nash bargaining solution (1950b). But the critical
idea was that this bargain would simply be a rational bargain: the only
normativity is the normativity of rationality.
Just as we had to deﬁne Alf’s utility function before examining what
a rational Alf would choose, contractarian approaches to moral and
political theory also require that we give content to utility functions.
Rational Choice Theory 235

Before showing that rationality leads to certain normative principles, we

must have some account of what people’s preferences are. Hobbes (1994:
ch. 11) does this when he posits in all mankind ‘a perpetual and restless
desire for power after power, that ceaseth only in death’. Gauthier (1986:
87) also does this when he insists that utility functions be non-tuistic, or
not other-regarding. This means that the satisfaction of one’s preferences
cannot rely on other people’s preferences being satisfied. Excluded by
non-tuism is the mother who only prefers to see her children’s preferences
satisfied. The reason Gauthier excludes such preferences is an attempt to
avoid his theory admitting solutions based on adaptive preference forma-
tion under domination: for example, Amartya Sen (2009: 166–7, 285–6)
notes how in some areas Indian women’s understandings of what they
want and how well they are doing are fundamentally shaped by their
acceptance of the normality of gender bias.
Contractarians such as Hobbes and Gauthier are the purest examples
of the project of building moral and political principles on the normativity
of rational choice. Rawls’s famous social contract theory also seeks to base
the normative appeal of its conclusions on the normativity of rational
choice. At the outset of A Theory of Justice Rawls (1999a: 15–16) insists
that
one conception of justice is more reasonable than another, or justifiable with
respect to it, if rational persons in the initial situation would choose its principles
over those of the other for the role of justice. Conceptions of justice are to be
ranked by their acceptability to persons so circumstanced. Understood in this way
the question of justification is settled by working out a problem of deliberation: we
have to ascertain which principles it would be rational to adopt given the con-
tractual situation. This connects the theory of justice with the theory of rational
choice.

And thus the importance of our opening quote from Rawls (1999a: 15):
‘The theory of justice is a part, perhaps the most signiﬁcant part, of the
theory of rational choice.’
Rawls, however, insists that the choice of moral principles is not simply
governed by the ‘rational’, but also the ‘reasonable’ – a concern with fair
cooperation (1993: 48ff). One way to model this is to select principles
behind a veil of ignorance, in which parties have drastically restricted
information sets, including information as to who they are, and what their
utility functions look like (Harsanyi 1955; Gaus and Thrasher 2015).
This restriction on information forces parties to choose fairly: hence non-
tuistic parties, i.e., choosers only concerned with maximizing their own
goals, are forced (via ignorance) to consider how the goals of each fare,
since once the veil is lifted, one may end up being any of these people.
236 Chapter 11

A rather more straightforward route, perhaps, is simply to build concerns

with fairness or morality into the utility functions of the choosers (Gaus
2011: ch. 5). Again, it is crucial to stress that rational choice theory does
not tell us what a person’s preferences are. The more one stresses the
‘reasonable’ over the ‘rational’, the more one is insisting that the selection
of principles must not only suppose utility maximizers, but those with
certain substantive moral concerns.

5 Decisions to Be Made When Using Decision Theory

We have tried to stress the complexities of rational choice theory, and
have tried to alert readers to the pitfalls of widespread misconceptions.
We can apply these lessons to thinking about how a researcher might
contemplate using a rational choice framework in her own work in
political theory.
It may sound terribly simple, but the first step is to be clear whether
you are using rational choice theory as an explanatory, predictive,
or normative model. These can be surprisingly easy to run together.
Think of Hobbesian political theory: part of Hobbes’s famous argument
is how people will act in the state of nature (an explanatory project) and
the next part is a normative contract: what would be a rational agreement
to leave the state of nature? These raise vastly different issues.
The explanatory part of the project has been advanced by game theoretic
modeling concerning the ease with which conflict can erupt in a state of
nature (Vanderschraaf 2006; Chung 2015). The specification of the
contract is normative: what constitutes a rational bargain? This leads to
the sorts of concerns Gauthier and Rawls have investigated. Confusion
between these two projects can lead to errors. In offering an empirical
analysis of the state of nature, Hobbesians are not saying that this is what
persons ought to do – in fact, the Hobbesian project is driven by a desire to
show how alleviating one’s self from such a state is consistent with rational
choice. And in offering normative analyses of the social contract, theorists
are not trying to describe anything that has happened, or necessarily will.
Instead, they are trying to model the reasons we have to obey certain
authority relations via a rational choice analysis.
Be clear about whether the problem is parametric or strategic.
Even the best political theorists can make this mistake. Rawls (1999a: 11)
said that we should understand the derivation of the theory of justice as
the result of a ‘fair agreement or bargain’, suggesting that he wanted to
model a strategic rather than parametric choice problem in the original
position. But later on Rawls (1999a: 120) insists that, due to the normal-
ization of the parties’ interests, we can really understand choice in the
Rational Choice Theory 237

original position ‘from the standpoint of one person selected at random’,

suggesting a parametric rather than strategic choice problem. This can
lead to confusion. Consider, for instance, Rawls’s (1999a: 132–3) con-
tinual insistence that we can think of the two principles of justice as ‘those
a person would choose for the design of a society in which his enemy is to
assign him his place.’ If Rawls models the original position as a parametric
choice problem, then such remarks conjure up the maximin choice
principle under uncertainty – which, as we have seen, is riddled with
controversy. But suppose Rawls models the original position as
a strategic choice problem rather than a parametric one. Then, if the
game is zero-sum, we have seen that the maximin choice is the uniquely
rational solution to the game, and thus not subject to controversy. Whether
the original position is parametric rather than strategic is thus highly
relevant in terms of its plausibility as a rational choice model. We end
up following the standard way of understanding the original position in
the literature as a parametric choice problem rather than a strategic choice
problem, though, again, Rawls’s language could be clearer on this point.
The really hard step is to determine the utility functions of the
parties. And how utility functions are defined can have tremendous
relevance for the plausibility of a rational choice model, regardless of
whether the model is normative or explanatory. Consider again
Gauthier’s normative contractarian project. As we mentioned, Gauthier
defines utility functions by requiring preferences be non-tuistic so that his
bargain is not the result of preference adaptation under domination. Such
stipulation, though, is not without controversy. Donald Hubin (1991)
argues that excluding tuistic preferences means that rational persons with
such preferences will not find it rational to comply with the resulting
bargain. After all, if the bargain was derived from a non-tuistic utility
function, but a rational person has a tuistic utility function, then it is hard
to see why it is rational for the other-regarding person to obey constraints
derived specifically for persons who are not other-regarding. And from
the explanatory side, Jean Hampton (1986: 75) argues that, given one
interpretation of how Hobbes defines utility functions of actors in the
state of nature, there would actually be no conflict in the state of nature at
all. If people really were rational actors who only cared only about self-
preservation, then the iterated context of the state of nature would allow
for cooperation. No need for the sovereign after all!
You should now decide whether to represent these preferences
ordinarily or cardinally. For some simple games (such as the
Prisoner’s Dilemma), an ordinal representation will be sufficient; but if
risk or uncertainty is involved, we have seen, a cardinal representation is
likely to be required.
238 Chapter 11

Your next decision is: what are the actor’s information sets?
Do they know everything, or only a little? And if one is modeling
strategic interaction, one must decide if what the parties know is the same,
or if some parties know more than others. This latter difference can be
crucial, again for both normative and explanatory rational choice models.
On the normative side, we again turn to contractarianism. One common
assumption in bargaining theory is the assumption that parties are ‘sym-
metric’ – that they have the same strategies available to them, the same
bargaining ability, and, importantly, the same information sets. This
assumption was originally introduced by Nash (1950b), but is common
in many bargaining models, including Gauthier (1986). Yet as John
Thrasher (2014) shows, this assumption can often be at odds with the
contractarian project – it amounts to importing fairness criteria into the
terms of the bargain, when really what the contractarian is trying to do is
derive fairness criteria from the bargain itself. For the reason the symmetry
assumption is often introduced is because once we admit things like
asymmetric information sets into the bargain, the resulting distribution
might not be so intuitively attractive when compared to the symmetrical
case. And from an explanatory standpoint, Zachary Ernst (2005) force-
fully argues that many explanatory models of the evolution of conventions
and fairness norms obscure more than they reveal when they assume
symmetry across parties, of which symmetrical information sets is one
such part. These sorts of considerations are highly relevant when deter-
mining the information sets of the parties – symmetrical information
might be clean and easy, but possibly problematic.
If you decide that you are concerned with a strategic choice, you
will need to learn a bit of game theory. There are a number of good
introductory texts (e.g., Maschler, Solan, and Zamir 2013). Do not be
too quick to focus on a particular game before you have
determined the preferences and the information sets of the
parties. Games are deﬁned by the preferences, information sets, and
strategies of the players. It is all too common for even postgraduate
students (and, alas, professors!) to get fascinated by a game such as the
Prisoner’s Dilemma or Chicken, and then to try to ﬁnd ways to apply it,
often forcing a case or situation into an inappropriate game. Once one has
made the previous decisions on our list, one can fruitfully think about
game theoretic representations.
A last lesson: when using rational choice theory, especially in
normative contexts, it is critical to realize its limits. Often our
disputes are not about what a rational person would do, but about the
correct preferences to ascribe to people, or the conditions for fair
bargains. Understanding moral and political principles as an object
Rational Choice Theory 239

of choice among rational choosers can clarify their basis, but we must
be very careful not to reduce all these to disputes about rationality.
Even experts can believe that their dispute is about rationality, rather
than the proper content of an agent’s concerns. A good example of
this is the Rawls-Harsanyi debate. Harsanyi (1953, 1955) and Rawls
pursued very similar contract-based normative projects, reducing the
problem of moral justification to one of rational choice behind the veil
of ignorance. Contra Rawls, however, Harsanyi believed that rational
parties choosing behind a veil of ignorance would select average
utilitarianism rather than Rawls’s two principles of justice. At first
blush the dispute between Rawls and Harsanyi reduced to an argu-
ment about rational choice under uncertainty. Rawls thought rational
parties, when faced with radical uncertainty, ought to adopt the max-
imin decision rule. Harsanyi thought rational parties, when faced with
radical uncertainty, ought to adopt the Laplacean principle of
indifference. The debate continued after both parties published their
seminal social contract contributions. Harsanyi (1974) continued to
point out the irrationality of the maximin decision rule, while Rawls
(1999b: 225–31) continued defending it.
Recent analysis of the Rawls-Harsanyi dispute, though, suggests that
the disagreement between the two cannot merely be reduced to a debate
about rational choice. Michael Moehler (2013) argues that Rawls’s and
Harsanyi’s disagreement over what rationality requires itself reduces to
a philosophical dispute about which moral values should be modeled in
the choice situation, and thus which moral values are important. Rawls
wanted to model impartiality in the original position. Harsanyi was
especially concerned with modeling impersonality. While impartiality
merely requires that agents do not unjustifiably favor their own interests
in making decisions, impersonality requires agents to make decisions only
on the basis of the common interest. This inclusion of impersonality,
which Rawls (1999a: 24) rejected, required Harsanyi to use the Laplacean
principle of indifference. And as Rawls (2001: 97–100) notes later on, the
use of maximin reasoning in the original position ultimately comes down
to a moral, not a rational choice. The putative debate between Rawls and
Harsanyi over the nature of rational choice theory reduces to
a disagreement over what is to be valued – impartiality versus imperson-
ality. This is an object lesson. Rational choice theory by no means
replaces philosophical analysis. Remember always that it is a tool,
and like all tools, is effective only when it is used to perform its
proper tasks in its proper ways.
240 Chapter 11

References
Amadae, S. M., 2015. Prisoners of Reason: Game Theory and Neoliberal Political
Economy. Cambridge University Press.
Aumann, Robert, 1987. ‘Correlated equilibrium as an expression of Bayesian
rationality’, Econometrica 55: 1–18.
Binmore, Ken, 2005. Natural Justice. Oxford University Press.
Binmore, Ken, 2009. Rational Decisions. Princeton, NJ: Princeton University
Press.
Bosman, Ronald, Matthias Sutter, and Frans van Winden, 2005. ‘The impact of
real effort and emotions in the power-to-take game’, Journal of Economic
Psychology 26, 407–29.
Buchanan, James and Gordon Tullock, 2004. The Calculus of Consent: The Logical
Foundations of Constitutional Democracy. Indianapolis, IN: Liberty Fund.
Chung, Hun, 2015. ‘Hobbes’s state of nature: a modern Bayesian game-theoretic
analysis’, Journal of the American Philosophical Association 1, 485–508.
Ernst, Zachary, 2005. ‘A plea for asymmetric games’, Journal of Philosophy 102:
109–25.
Friedman, Jeffrey, ed., 1996. The Rational Choice Controversy: Economic Models of
Politics Reconsidered. New Haven, CT: Yale University Press.
Gaus, Gerald, 2011. The Order of Public Reason. Cambridge: Cambridge University
Press.
Gaus, Gerald and John Thrasher, 2015. ‘Rational choice and the original position:
the (many) models of Rawls and Harsanyi’, in Timothy Hinton, ed., The Original
Position. Cambridge, MA: Harvard University Press, 39–58.
Gauthier, David, 1986. Morals by Agreement. Oxford: Clarendon Press.
Gauthier, David, 1991. ‘Why contractarianism?’, in Peter Vallentyne, ed.,
Contractarianism and Rational Choice. Cambridge University Press, 15–30.
Gintis, Herbert, 2009. The Bounds of Reasons: Game Theory and the Uniﬁcation of the
Behavioral Sciences. Princeton, NJ: Princeton University Press.
Hamilton, W. D., 1964. ‘The genetical evolution of social behaviour I’, Journal of
Theoretical Biology, 7: 1–16.
Hampton, Jean, 1986. Hobbes and the Social Contract Tradition. Cambridge
University Press.
Harsanyi, John, 1953. ‘Cardinal utility in welfare economics and in the theory of
risk-taking’, Journal of Political Economy 61: 434–5.
Harsanyi, John, 1955. ‘Cardinal welfare, individualistic ethics and interpersonal
comparisons of utility’, Journal of Political Economy 63: 309–21.
Harsanyi, John, 1974. ‘Can the maximin principle serve as the basis for morality?’,
American Political Science Review 69: 594–606.
Harsanyi, John and Reinhard Selten, 1988. A General Theory of Equilibrium
Selection in Games. Cambridge, MA: MIT Press.
Hobbes, Thomas, 1994. Leviathan, ed. Edwin Curley. Indianapolis, IN: Hackett
Publishing.
Hubin, Donald, 1991. ‘Non-tuism’, Canadian Journal of Philosophy 21: 441–68.
Rational Choice Theory 241

Kalai, Ehud and Meir Smorodinsky, 1975. ‘Other solutions to Nash’s bargaining
problem’, Econometrica 43: 513–18.
Kliemt, Hartmut, 2009. Philosophy and Economics I: Methods and Models. Munich:
Oldenbourg.
Luce, Duncan and Howard Raiffa, 1957. Games and Decisions. New York: Dover
Publications.
Maschler, Michael, Eilon Solan, and Shmuel Zamir, 2013. Game Theory.
Cambridge University Press.
Moehler, Michael, 2013. ‘Contractarian ethics and Harsanyi’s two justifications of
utilitarianism’, Politics, Philosophy, and Economics 12: 24–47.
Nash, John, 1950a. ‘Equilibrium points in n-person games’, Proceedings of the
National Academy of Sciences of the United States of America 36: 48–49.
Nash, John, 1950b. ‘The bargaining problem’, Econometrica 18: 155–62.
Nash, John, 1951. ‘Non-cooperative games’, Annals of Mathematics 54: 286–95.
Plamenatz, John, 1973. Democracy and Illusion. London: Longman.
Plott, Charles, 1996. ‘Rational choice in experimental markets’, in Robin Hogarth
and Melvin Reder, eds., Rational Choice: The Contrast between Economics and
Psychology. Chicago, IL: University of Chicago Press, 117–43.
Rawls, John, 1993. Political Liberalism. New York: Columbia University Press.
Rawls, John, 1999a. A Theory of Justice, Revised Edition. Cambridge, MA:
Harvard University Press.
Rawls, John, 1999b. Collected Papers. Cambridge, MA: Harvard University Press.
Rawls, John, 2001. Justice as Fairness: A Restatement. Cambridge, MA: Harvard
University Press.
Rosenberg, Alexander, 1995. Philosophy of Social Science, 2nd edition. Boulder,
CO: Westview.
Rubinstein, Ariel, 1982. ‘Perfect equilibrium in a bargaining model’, Econometrica
50, 97–110.
Schelling, Thomas, 1960. The Strategy of Conflict. Cambridge, MA: Harvard
University Press.
Schwartz, Norbert, 2002. ‘Feelings as information: moods influence judgments
and processing strategies’, in Thomas Gilovich, Dale Griffin, and
Daniel Kahneman, eds., Heuristics and Biases: The Psychology of Intuitive
Judgments. Cambridge University Press, 534–47.
Sen, Amartya, 2009. The Idea of Justice. Cambridge, MA: Harvard University
Press.
Shapiro, Ian and Donald Green, 1996. Pathologies of Rational Choice: A Critique of
Applications in Political Science. New Haven, CT: Yale University Press.
Skyrms, Brian, 1996. Evolution of the Social Contract. Cambridge University Press.
Skyrms, Brian, 2004. The Stag Hunt and the Evolution of Social Structure.
Cambridge University Press.
Smith, John Maynard, 1975. The Theory of Evolution, 3rd edition. New York:
Penguin.
Smith, Vernon, 2002. Bargaining and Market Behavior: Essays in Experimental
Economics. Cambridge University Press.
242 Chapter 11

Sugden, Robert, 1986. The Economics of Rights, Cooperation and Welfare. Oxford:
Blackwell.
Thrasher, John, 2014. ‘Uniqueness and symmetry in bargaining theories of
justice’, Philosophical Studies 167: 683–99.
Thrasher, John and Gerald Gaus, forthcoming. ‘James Buchanan and Gordon
Tullock, Calculus of Consent’, in Jacob Levy, ed., The Oxford Handbook of Classics
in Contemporary Political Theory. Oxford University Press.
Trivers, Robert, 1971. ‘The evolution of reciprocal altruism’, The Quarterly Review
of Biology 66: 35–57.
Vanderschraaf, Peter, 2006. ‘War or peace? A dynamical analysis of anarchy’,
Economics and Philosophy 22: 243–79.
von Neumann, John, 1928. ‘Zur Theorie der Gesellschaftsspiele’, Mathematische
Annalen 100: 295–320.
von Neumann, John and Oscar Morgenstern, 1944. The Theory of Games and
Economic Behavior. Princeton, NJ: Princeton University Press.
Weisberg, Michael, 2007. ‘Who is a modeler?’, British Journal for Philosophy of
Science 58, 207–33.
12 Interpreting Texts

Adrian Blau

1 Introduction: Meanings and Understandings

There are three main ways of interpreting texts, based on three kinds of
understanding and three equivalent kinds of meaning: (1) what authors
mean, which is empirical; (2) what the ideas mean, which is philosophical;
and (3) what one or both of these mean to the reader – how you or I feel
when we read a text – which is aesthetic.
Aesthetic interpretation is primarily the province of literature depart-
ments. Of course, what texts mean to those of us in other departments can
still influence what we study. Rousseau has made me cry, because of the
beauty and passion of his writing. Habermas has made me cry, for other
reasons. Such reactions can influence our empirical and philosophical
understanding. And we do sometimes study the effects authors had or
tried to have on their readers, for example through their rhetoric. But
what texts mean to us is not usually our academic focus, and I discuss it no
further.
By contrast, philosophical interpretation – what the ideas mean – is
very significant for us. Not that you would know it from our
methodological literature, which mainly emphasizes empirical issues,
especially what authors mean. Yet both kinds of meaning matter when
interpreting texts, because they involve different kinds of under-
standing. If you read Mill’s On Liberty and understand exactly what
he meant by every word he used, you have understood something very
important. But you understand his writing better if you also spot his
ambiguities, contradictions, successes and failures. Unfortunately, the
best methodological writings about textual interpretation – those by
Quentin Skinner (2002a) – say almost nothing about the second kind
of meaning and understanding, and hence imply that the first can
always be achieved on its own.
I seek to connect these two kinds of meaning and these two kinds of
understanding. A piece of research typically prioritizes one of the two, but
almost always handles both. They are not alternatives: we usually need

243
244 Chapter 12

one to find the other. That fundamental point has not, I believe, been
made in previous methodological discussions.
Equally unfortunately, most methodological writings give the wrong
impression by talking about different ‘approaches’ or ‘schools of thought’,
like contextualism, Straussianism, Marxism and so on. These categories
have some value, and I cover them later. But mental categories can limit
our thoughts and actions, and I believe this has happened here: most
commentators imply that these categories are very different without add-
ing that there are principles of good interpretation that apply to all of us
(see especially Pocock 1971: 6–11; Rorty 1984: 49; Dunn 1996: 19; Ball
2004: 19; Richter 2009: 7–11; Schulz and Weiss 2010: 284–8). Despite
some marked differences in our interpretive aims, there is or should be
much overlap in our interpretive methods. Silence about these core
principles has not prevented excellent interpretations, but has permitted
more dubious scholarship than otherwise.
This chapter’s main aim is to make explicit these core principles of
good interpretation. You will find relevant principles in every section of
this chapter, even in sections you might think do not apply to you.
Three brief caveats. First, my examples are mostly historical, but the
principles also fit contemporary texts. Second, my examples mainly come
from well-known Western authors, like Machiavelli and Rousseau, but
the principles also apply to less well-known Western authors, and to
authors in other cultures. (See also the chapters in this volume by
Leader Maynard on ideological analysis, and by Ackerly and Bajpai on
comparative political thought.) Third, and most important, although
I cover both empirical interpretation (e.g. what Locke meant by ‘prop-
erty’) and philosophical interpretation (e.g. how well Locke’s defence of
property works), and although I link the two more than other commenta-
tors do, most of my how-to guidance involves empirical interpretation,
including using philosophical interpretation as part of empirical
interpretation. Other chapters in this volume will help more for readers
primarily interested in philosophical interpretation in itself (especially the
chapters by Knight on reflective equilibrium, and Brownlee and
Stemplowska on thought experiments). But you should still read this
chapter to the extent that you want to get historical or contemporary
authors right. And Olsthoorn’s chapter (especially Section 6) will help for
both empirical and philosophical interpretation.
Section 2 summarizes contextualism: despite its crucially important
focus on history and context, there are other secrets to contextualists’
success, and pitfalls we must all beware. Section 3 addresses
Begriffsgeschichte, conceptual history and genealogy, which combine tex-
tual and/or contextual analysis with conceptual comparison. Section 4,
Interpreting Texts 245

the most important section, considers reconstruction, which is sometimes

seen as purely philosophical, but which we all do – even historians.
Section 5 questions the usefulness of treating perspectives like feminism
or Marxism as if they are ‘approaches’. Rather, they provide hypotheses
and distinctions that help us see things that other scholars overlook, and
these ‘approaches’ are far more numerous than we usually think. Nor is
‘reading between the lines’ exclusively a Straussian tool, Section 6 argues.
Section 7 then offers further general principles of good interpretation that
apply to all of us – whatever categories, approaches or schools of thought
we identify with.

2 Contextualism
Contextualism is often called the ‘Cambridge School’ approach, albeit
misleadingly (Skinner 2012: 16–17). Contextualism has dominated
methodological debates about textual interpretation for half a century.
And contextualists have essentially won the battle, both through their
theoretical arguments and the high quality of actual contextual research.
Although their theoretical arguments sometimes remain controversial,
the core idea is widely accepted: place texts in their historical contexts.
Even scholars who do not themselves do historical research usually
consult such publications, to make misinterpretation less likely.
Contextualism has a long history (e.g. Allen 1928: xvii–xx), but started
its upsurge in the late 1940s and flowered in the 1960s, with the theore-
tical ideas and substantive interpretations of writers like Peter Laslett,
John Pocock, Quentin Skinner and John Dunn (Pocock 2006: 37–9).
Given diversity within and between these writers and their followers
(Boucher 1985: 151–272; Skinner 2012: 16–17), I focus mainly on
Skinner, who I see as the supreme methodologist in our field. I have
learned much from his methodological writings, but even more from his
substantive interpretations, which I find methodologically richer and
more impressive. By contrast, Mark Bevir only tackles Skinner’s metho-
dological writings (1999: 40–50, 327) and thus overlooks some of
Skinner’s most important methodological lessons (see also Stuurman
2000: 319; Skinner 2002a: 178–9). Treating Skinner as a practitioner,
not just a theorist, lets me sidestep his speech-act theorizing, which is
essentially separate to contextual analysis (Hutton 2014: 927), and which
I address elsewhere.1

1
Adrian Blau, ‘Extended meaning and understanding in the history of ideas’ (working
paper, 2016).
246 Chapter 12

Although Skinner does not say so, and although most commentators
emphasize his contextualism, the foundation of Skinner’s success is
actually close textual analysis. Indeed, he has taught history of political
thought to graduate students like this, such as eight weeks of close reading
of Leviathan in Cambridge. The first principle of contextualism, and all
sensible interpretation, is thus: read texts carefully.
This requires us to read passages in their textual contexts.
Consider Hobbes’s comment that ‘the Thoughts, are to the Desires, as
Scouts, and Spies, to range abroad, and find the way to the things desired’
(Hobbes 1991: 8.16, 53). Although this does not mention ‘reason’, many
scholars seem to equate ‘thoughts’ and ‘reason’, and treat Hobbes as
implying that reason is instrumental to the passions, or even the slave of
the passions. But in the context of the chapter as a whole, this comment is
surely not about reason: Hobbes is merely saying that our thoughts jump
to what we desire. If I like baseball and see a big stick, my mind might
think: ‘could I hit a baseball with that?’ (Blau 2016: 201–2).
We cannot always read much of an author’s output: inevitably, we
sometimes dip in, especially for authors who are not central to our
work. Being corrected on such grounds is an occupational hazard. But
ideally, we should read an author’s texts widely, to avoid overlooking
important passages elsewhere in the text and in other texts. Book 2
chapter 3 of Rousseau’s Social Contract makes more sense alongside
book 4 chapter 1, casting light on how Rousseau may have understood
the general will (which is not to say that this is what Rousseau had in
mind). ‘Non-political’ texts might also be relevant: Rousseau’s Emile,
ostensibly about education, might also be read into these passages (e.g.
Rousseau 1979a: book 4, p. 286): perhaps morally pure people can look
inside themselves and intuit the general will. That said, the draft of the
Social Contract implies that intuitionism is inappropriate in modern
society (Rousseau 1997a: 1.2.6–14, 154–7). Yet the exclusion of these
passages from the final version must give us pause for thought. You never
know when to read different ideas/texts into each other and when
they are not consistent. Consider different options.
Textual analysis alone is never enough: a key contextualist contribution
is to place texts in their linguistic contexts by reading other texts
from the similar time/place, or by reading the work of scholars
who have done this. This can let us understand words we no longer use,
like dehort (dissuade). It helps us spot ‘false friends’ – words that look
familiar but whose meanings once differed, like prejudice (pre-judgement,
not bias), or force, which in eighteenth-century French had a sense of
strength/energy that might not be recoverable from merely textual
analysis of Rousseau’s ‘forced to be free’ (Mason 1995: 135–6).
Interpreting Texts 247

Placing texts in their linguistic contexts also helps us infer

intentions and indicates originality. Machiavelli uses fortuna
conventionally but virtú unconventionally, dramatically undercutting
the orthodoxies of his day (Skinner 1981: 24–31, 34–47). If we only
read the ‘canon’ of great thinkers – Aristotle, Locke, Marx and others –
we will miss part of what made these texts so ground-breaking. For some
scholars, contextualism has had ‘a hugely negative impact’ by making the
study of past thinkers ‘merely antiquarian’, even ‘frivolous’ (Kelly 2006:
48). But it can be wonderfully exciting to see how authors interacted with,
built on or subtly knifed their lesser-known contemporaries, who can be
wonderfully exciting in themselves.
We should thus place texts in their intellectual contexts – politi-
cal, philosophical and so on. We can understand much of Hobbes’s
Leviathan when simply read as an abstract account of citizens’ rights and
duties, but when we also see it in its civil war political context and its anti-
Aristotelian philosophical context, we understand more of it, we under-
stand parts of it differently, and we understand parts of it better. That
philosophical context, note, involves both Aristotle himself and Hobbes’s
contemporaries, whose use of Aristotle Hobbes saw as dangerous, indeed
as one cause of the civil war: Hobbes’s political and philosophical con-
texts overlap. Incidentally, Hobbes’s ‘Aristotle’ is not our ‘Aristotle’: we
may need to consult the 1598 English translation of Aristotle’s Politics that
Hobbes used, not a modern edition.
Placing texts in their contexts may not solve our problems. For
example, contextualism shows what notions of virtue Machiavelli was
probably challenging and why, but to uncover his own conception(s), we
must think through his arguments and examples philosophically, and see
what ﬁts. Section 4 discusses this in more detail. Indeed, this chapter
frequently stresses the need to combine empirical and philosophical
analysis. Contextualists often do this brilliantly, but it is not covered in
their methodological writings. For example, Dunn’s (1967) analysis of
Locke on consent powerfully combines historical and philosophical rea-
soning; but in his methodological discussion of the interdependency of
historical and philosophical interpretation (1968: 85–8), he does not say
that to uncover what authors mean, we may need to analyse their argu-
ments philosophically, not just contextually. Worse, some historians
explicitly disdain philosophical analysis (e.g. Laslett 1988: 81–90; for
criticisms of Laslett, see Waldron 2002: 50–2, 62, 68, 190–1).
Even when contexts give an answer about authors’ meanings or
motives, do not assume that the answer is complete. Consider
Rousseau’s claim that a man can be forced to obey the law and still be
free (Rousseau 1997a: 1.7.8, 53). Rousseau’s justiﬁcation is unclear.
248 Chapter 12

Helena Rosenblatt suggests that Rousseau, like many opponents of the

Genevan government, implies a traditional Christian/Calvinist view of
freedom. ‘Just like abiding by God’s will makes men free in Christian
thought, abiding by the general will makes citizens free in Rousseau’s
thought’ (Rosenblatt 1997: 255–6; also 246–7). This explanation looks
right, but it is not enough: it may not fit Rousseau’s ensuing comment that
forcing someone to obey the general will makes him free by protecting
him from dependence on other people. Why that should be so requires us
to think through Rousseau’s ideas philosophically.
Skinner does this kind of philosophizing expertly in arguing that
Hobbes changes his account of liberty for contextual and philosophical
reasons (2008: 24, 45, 108, 132–8). But Skinner underplays the political-
economic and theological contexts, which imply somewhat different con-
clusions (Whatmore 2006: 121–5). More than one context may be
relevant. Note, indeed, that I have been referring to contexts, plural.
Remember that historical parallels can be coincidental.
According to Richard Tuck, Hobbes was responding to a form of scepti-
cism (1993: 285–7, 293–8, 304–7, 316). Most commentators, though,
explain the parallels differently (see Zagorin 1993: 512–18 for more
details). No evidence, including contextual evidence, is conclusive: the
same evidence can always be read differently.
Such points have not been adequately theorized by contextualists, who
have trumpeted the value of historical interpretation without saying much
about how to do it well (Green 2015: 436). The starting point, as
Section 7 argues, is to place hypotheses, inference and evidence centre
stage. Historians who doubt this should consider what happens when one
‘knows’ the answer and just seeks evidence that fits it, as with Leo Strauss
(Blau 2012).
Contextualization can even help with recent authors. We understand
Rawls better by placing him in his philosophical and political contexts
(see, respectively, Wolff 2013; Forrester 2014). But we can still under-
stand much of Rawls acontextually – just as with some historical passages.
Contextual analysis usually helps, but it is not always necessary,
and is often insufficient.
I have treated contextualism as seeking to recover authors’ meanings
and motives, through textual, contextual and philosophical analysis
(although the last of these is rarely noted). But it is unfair to complain
that ‘contextualist methodology . . . reduces the point of political theory to
historical questions of authorial intention and meaning’ (Kelly 2007: 11).
For some historians and many political theorists and philosophers, reco-
vering authorial intentions and meanings is at least partly aimed at other
things, including normative arguments, as we will now see.
Interpreting Texts 249

3 Begriffsgeschichte, Conceptual History and Genealogy

Begriffsgeschichte, the ‘history of concepts’, is usually associated with
Reinhart Koselleck, who co-led the Geschichtliche Grundbegriffe (in
English: Basic Concepts in History), a multi-volume, 7,000-page analysis
of more than 100 social and political concepts, published in German
between 1972 and 1997. This enterprise was so big, requiring so many
years and so many authors, that I do not cover it here. (For more
information, see Olsen 2012: chapter 4. For much more detail on the
Geschichtliche Grundbegriffe and Begriffsgeschichte, see Richter 1995,
especially pp. 124–42 on similarities and differences between
Begriffsgeschichte and contextualism.)
I focus instead on the smaller-scale version of Begriffsgeschichte: con-
ceptual history, or genealogy. I treat the terms as equivalent: Skinner now
calls his conceptual history of the state a genealogy of the state (Skinner
1989, 2009). Genealogy, a Rousseauian term (Griswold 2016: 277), is for
most scholars associated with Nietzsche, whose genealogical approach is
actually wider than that of historians like Skinner (Lane 2012: 75–82).
(For a comparison of the ideas of genealogy in Nietzsche and Foucault,
see Prado 2006: 76–81.)
Conceptual history has long been practised informally, but it arrived
self-consciously in the Anglo-American mainstream with Terence Ball,
James Farr and Russell Hanson’s edited book Political Innovation and
Conceptual Change (1989). Each chapter takes a concept, like corruption
or patriotism, and examines different conceptions of that concept over
time and sometimes place. (For the concept/conception distinction – i.e.
a general idea, and particular versions of that idea – see Rawls 1971: 5,
and Olsthoorn’s chapter in this volume, Section 2.2.) For example, Peter
Euben (1989) discusses changing senses of ‘corruption’, from physical
decay, through the lack of civic virtue in the body politic, to the much
narrower notion that dominates today. (For a partial critique of Euben,
see Blau 2009: 614–16. For a book-length conceptual history of corrup-
tion, see Buchan and Hill 2014.)
Conceptual history has two parts: primarily empirical and primarily
conceptual. The primarily empirical part will typically include textual,
contextual and philosophical analysis; the aim is to recover authors’ own
understandings of their terms (Skinner 2002a: 50). But because the aim is
comparison, rather than understanding a single author, case selection
needs extra attention. The ideal – taking all possible cases – is impractical,
especially in a single chapter or article. Be conscious, and perhaps
explicit, about how your case selection may affect your conclu-
sions, such as whether your claims are limited to a particular time, place
250 Chapter 12

or language (e.g. Skinner 2009: 325). You may need to consider words
not used, as well as words used, as with Josiah Ober’s analysis of
‘democracy’ and similar terms in ancient Greece (2008: 5, 7). And
distinguish the word from the idea (see Olsthoorn’s chapter in this
volume, Section 2.1).
The primarily conceptual part of the analysis involves conceptual
comparison: we stand back and compare authors’ understandings, e.g.
asking how similarly Bentham and Mill saw ‘utility’. Careful
conceptual analysis is vital, as with Skinner’s fine-grained genealogy
of liberty (2003: 22). By contrast, Isaiah Berlin’s (1969) confused and
confusing conceptual analysis undermines his conceptual history of
liberty (for a critique, see Cohen 1960). Conceptual history is not
just history: it is also conceptual – philosophical. Olsthoorn’s chapter
in this volume has many other tips to help rigorous conceptual analysis
in conceptual history.
Anachronistic conceptualizations may help, especially when com-
paring authors who used different terms. Anachronism is dangerous: it
can infect our efforts to recover authors’ meanings (Skinner 2002a:
49–51). For example, James Mill’s ‘middle rank’ does not mean
‘middle class’ (Ball 1992: xx–xxi). But if handled carefully, using
appropriately fine-grained distinctions, and preferably not until you
have first recovered authors’ meanings, anachronism lets us apply
conceptual frameworks that usefully highlight similarities and
differences between authors. For example, Rousseau’s Considerations
on the Government of Poland recommends that citizens elect deputies
every six weeks, on an explicit set of instructions, divergence from
which would see deputies being executed (Rousseau 1997b: 7.14–19,
201–3). This is an extreme example of the so-called mandate or
delegate conception of representation (Pitkin 1967: 145–7).
Rousseau did not use these terms: he even denies that this is ‘repre-
sentation’ (1997a: 3.15.5–6, 114). But what he says amounts to how
we use these terms. If we are to compare notions of representation
over time, we will almost certainly have to apply such anachronisms.
So, first try to work out what authors meant, then see how well
this fits your own conceptualization or an existing one.
Conceptual history is interesting and important, not only histori-
cally and politically, but also normatively: we see why some
conceptions won out, how new and old conceptions often differ, and
how old conceptions might revive contemporary discussions, as with
Skinner’s revealing conceptual histories of freedom and the state
(Skinner 2003, 2009). Be attentive to the normative implications
of your interpretations.
Interpreting Texts 251

4 Reconstruction
Reconstruction means testing and potentially supplying, supplementing,
modifying or removing presuppositions, deﬁnitions, links between com-
ments/ideas and steps in arguments, as with Skinner’s analysis of Hobbes
on representation (2002b: 177–208), John Gray’s account of Mill on
happiness (1996: 70–85) and A. J. Simmons’s testing of Locke’s moral
theory (1992: 14–67). Note that these scholars are, respectively,
a historian, a political theorist and a philosopher.
Reconstruction is often seen as a special, philosophical technique (e.g.
Rorty 1984: 49–53), but everyone reconstructs, simply to understand
what is said or written. Some degree and some kind of reconstruction
are inseparable from the very nature of interpretation, and necessary to
the understanding of our texts. Again, compartmentalizing approaches
into different categories has led us to overlook key unifying ideas.
Simplifying somewhat, I distinguish between three kinds of
reconstruction:
(a) empirical reconstruction – trying to work out what authors meant;
(b) systematic reconstruction – linking authors’ ideas, making implicit
distinctions explicit, assessing consistency and so on, whether or not
authors themselves saw these things;
(c) adaptive reconstruction – altering what authors wrote and perhaps
what they intended.
These are not alternatives. We almost always do all three simultaneously,
to greater or lesser extents. Even the simplest empirical reconstructions
add, subtract or amend words. Even the most extreme adaptive recon-
structions make some assumptions about what authors meant. And we
typically interpret authors with some kind of systematic reconstruction,
e.g. reading one idea into another even when the author does not specify
a link. (Strictly speaking, systematic reconstruction is not a separate
category, but a mix of empirical and adaptive reconstruction where we
might be agnostic about the extent to which authors were conscious of the
systematization we offer.)
Some historians, including Skinner (1964), worry about systematic
and adaptive reconstruction. My point is that we cannot avoid doing
either, so we should learn to do them well – as, indeed, with Skinner’s
virtuoso systematic reconstruction of Hobbes’s views on the state and
representation (2002b: 177–208), including carefully worded adapta-
tions of Hobbes’s views (2002c: 190; see likewise 2002b: 217, 235).
A key message of this section, and of this chapter, is thus that thinking
philosophically helps us reconstruct what authors meant, even if some
scholars baulk at altering or ‘improving’ the authors they study.
252 Chapter 12

I will start with the most minimal kinds of reconstruction that we all do,
every day. All communication involves resolving ambiguities and filling in
gaps: we cannot say everything we want to say (Searle 1978). If you ask,
‘Did you have a good day?’, you presumably mean, ‘Did you have a
good day [today]?’, not, ‘Did you have a good day [five weeks after you
were born]?’. Likewise, when Sidgwick (1981: ix) writes that he has made
‘numerous alterations and additions’ in preparing the second edition of
The Methods of Ethics, he presumably means ‘numerous alterations and
additions [to this book]’, not ‘numerous alterations and additions [to my
trousers]’, even though that is logically consistent with what he wrote, and
might have been a better use of his time.
During conversations, our brains reconstruct such comments in a flash.
When we read, the process is often slower and more conscious. For
example, Mill writes: ‘As it is useful that while mankind are imperfect
there should be different opinions, so is it that there should be different
experiments of living’ (1989: 3.1, p. 57). Is Mill commending experi-
ments of living only while mankind is imperfect, or even once perfected?
The latter, it turns out, fits better with what Mill later says about indivi-
duality and self-development. Of course, ‘fits better’ does not signify that
he meant this, or even that he was clear in his own head. We can easily
over-interpret authors and make them more consistent than they were: no
author is fully consistent (Skinner 2002a: 67–72). But thinking through
the implications of ideas, and probing their consistency with other
things they said or implied, gives us evidence about what authors
might have meant. Contextual analysis is all well and good – and it is
very, very good – but it goes hand in hand with more philosophical
analysis, even merely of the kind that probes consistency.
Another use of consistency involves testing authors’ definitions, or
testing possible definitions where authors do not provide them. For
example, Machiavelli never defines virtú, and contextual analysis only
tells us what Machiavelli was reacting against, not how he understood the
term. It takes a more philosophical cast of mind to stand back from
the text and reconstruct what Machiavelli means when he discusses
virtú: he generally means ‘those qualities, whether moral or
otherwise . . . most conducive to military and political success’ (Skinner
2002a: 48).
This is not to imply that a single notion must fit a term. C. G. Prado
(2006: 81–103) finds five kinds of truth in Foucault’s many writings.
It does not really matter how aware Foucault was of this: these ideas are
there, and we understand Foucault better by systematizing his comments
more neatly. Indeed, we should do this even if Foucault emphatically
insisted that he only had one notion of truth. We never fully control what
Interpreting Texts 253

we say or write. Mill’s ‘one very simple principle’ (1989: 1.9, 13) turns out
to be neither one nor very simple. We understand Mill better by distin-
guishing between what he says he will do, and what he does do. We should
be careful about taking authors at face value: to say that we must
recover authors’ motives in order to understand what they mean (Skinner
2002a: 103–7) should not mean that we are bound by those intentions
(Skinner 2002a: 110–11).
A more complex example involves an apparent error in On Liberty. Mill
emphatically argues that you can do as you like, even harm yourself or
cause offence, provided you do not harm others (especially 1989: 1.9, 13;
2.44, 54–5). But suddenly Mill says, briefly and without justification, that
you cannot do self-harming things in public if they cause offence (1989:
5.7, 98). This seems like an astonishing about-turn. To understand what
actions Mill thought permissible, we must think through these passages
and see if this apparent contradiction is resolvable. This could imply
a different notion of harm or offence than we had assumed, potentially
entailing a significant reinterpretation. If we cannot find such
a resolution, we might want to disregard this passing comment (as
recommended by Gray 1996: 102). We might still explain why Mill
wrote it, e.g. to protect himself from objections of libertinism, but we
risk misunderstanding Mill if we treat this apparently incidental passage
as his considered view of harm. When you find an apparent contra-
diction, consider how (if at all) to resolve it – including the possi-
bility that it is not a contradiction at all.
A still more complex example involves Hobbes’s ambiguous references
in De Cive to ‘the dictates of right reason’ (1998: e.g. 2.2, 34; 15.4, 173).
Does he mean that reason forces our passions to obey, or merely that it
tells us what we should do even if our passions then disregard this? Or has
he not fully thought this issue through? (See Blau 2016.)
In effect, simply to understand one phrase we may need to reconstruct
Hobbes’s account of reason, the passions, deliberation and their
interconnections – quite an enterprise. It is thus problematic that con-
textualists encourage us to place texts in their contexts without addressing
the value of thinking through texts philosophically. Both matter.
This example actually has two even trickier dimensions. First, although
Hobbes’s other writings might help us interpret these comments in De
Cive, his account of reason might have changed over time (Skinner 1996:
298–437). Understanding the De Cive passages may involve tackling this
much bigger issue. Second, and even worse, our judgement of the change-
over-time thesis itself rests partly on how we read the comments about
reason and the passions in De Cive and elsewhere. In other words, the
thesis that informs our interpretation of individual comments itself
254 Chapter 12

depends on how we read those comments. Many scholars describe such

situations as ‘hermeneutic circles’. I argue elsewhere that the language
and literature of hypotheses is more precise and practical than the lan-
guage and literature of hermeneutics (Blau 2015b). We read Hobbes’s
comments, alongside contextual evidence, to see which hypothesis – that
Hobbes’s account changed, or stayed fairly constant – best fits the
evidence, what evidence does not fit or remains unclear and what this
implies for the hypotheses overall. (There are other hypotheses, of course,
including Hobbes being unclear or confused. See Blau 2016: 213–14.)
Such analysis is never conclusive, but no empirical analysis is conclusive.
The best approach is to be aware of interdependent interpretations
and consider different hypotheses.
Our assessment of authors’ consistency helps our interpretations. For
example, Rousseau states that freedom is linked to the general will, but
leaves the nature of the link unclear (e.g. 1997a: 1.7.8, 53; 4.2.8, 124).
We are more likely to reconstruct the nature of this link in a way that also
fits Rousseau’s other ideas if we read him as a rigorous, systematic
philosopher than if we read him as a polemicist focusing more on politics
than philosophy, or an erratic rhetorician who would say anything that
sounded good. As with Hobbes’s alleged change over time, these inter-
pretations are interdependent: whether we find a workable link between
freedom and the general will may affect how hard we seek links in other
places in Rousseau. Again, understanding even one or a few com-
ments may require more systematic reconstruction.
Sometimes such systematic reconstructions are a key aim, rather than
merely a means to uncover meaning. Consider John Gray’s (1995) sys-
tematic reconstruction of Isaiah Berlin’s liberalism. He collects Berlin’s
arguments in one place, clarifies key ideas (e.g. freedom) and shows their
links and underpinnings, such as Berlin’s agonism and value-pluralism.
Gray tries to stay true to Berlin, but is clearer and more compelling.
To the extent that Gray’s reconstruction is right, we understand Berlin
better than by reading Berlin alone. Gray’s reconstruction also lets him
identify weaknesses in Berlin’s position, resolving which might require
significant changes (1995: 2, 143, 151–68). To the extent that this
adaptive reconstruction is right, we understand Berlin even better.
I now address another book-length reconstruction – A. J. Simmons’s
(1992) analysis of Locke, revered by many political theorists and philo-
sophers, but largely sidestepped by historians. One reason for this neglect,
I fear, is that while historians have successfully made the case that political
theorists and philosophers should address historical research, political
theorists and philosophers have not made the reverse case so effectively.
Another reason may be that historians dislike Simmons’s adaptive
Interpreting Texts 255

reconstruction of Locke for contemporary purposes. But Simmons also

does empirical and systematic reconstruction. And his techniques are
often brilliant, worthy of contemplation by even the most empirical of
historians, because they can help us recover the beliefs of authors we
study, especially more philosophical authors.
Much of Simmons’s reconstruction is fairly standard, filling gaps and
resolving ambiguities by using textual, contextual and/or simple philoso-
phical reasoning. For example, when trying to understand what Locke
meant about making something one’s property by mixing one’s labour
with it, Simmons defends his interpretation over Tully’s by pointing to
how Locke used certain words, by asking whether certain arguments
really were historically unavailable to Locke and by looking at the impli-
cations of ideas: ‘it is awfully hard to believe that Locke thought picking
up an acorn to be sufficiently like God’s act of creation to generate rights
under the same creationist principle’ inferred by Tully (Simmons 1992:
252–64; but compare Tully 1995: 114–19).
Such philosophical reasoning is straightforward. Importantly,
though, Simmons’s empirical and systematic reconstructions some-
times involve greater philosophical sophistication. For instance, he
chooses to separate out the different steps in an argument of
Locke’s (Simmons 1992: 23) – a powerful technique that helps us
probe arguments’ strengths and weaknesses, missing steps, unstated
assumptions and so on (see Simmons 1992: 23–36, 46). Locke does
not flesh out the final step, for example, but Simmons uses what
Locke wrote to get ‘a good sense’ of what duties Locke might have
thought we had (1992: 59–67).
Another helpful technique is to list the possible or plausible mean-
ings of ambiguous words or ideas (Blau 2015c: 1184–5; 2016:
213–14). This technique helps Simmons think through what Locke
might have meant by such vague but important comments as the ‘original
community of all things’ and ‘enough, and as good’ (1992: 237–40,
295–8). More controversially, to try to grasp what Locke might have
meant by ‘rights’, Simmons anachronistically applies modern
distinctions, from writers like Hohfeld, Hart and others (1992: 70–5,
85–94). An even more controversial anachronism is Simmons’s rule-
consequentialist interpretation of Locke’s moral theory that ‘fit[s] . . .
Locke’s most firmly entrenched positions’ (but not everything he
wrote), lets us ‘make sense of some central but puzzling passages in
the Second Treatise’, and is thus Simmons’s ‘best guess’ at Locke’s
‘undoubtedly inchoate’ intentions (1992: 46–59). Anachronistic ana-
lysis is dangerous and needs considerable care, but it can
256 Chapter 12

potentially help us see what authors meant or were getting at, by

showing us interpretive possibilities that we might have missed.
Perhaps unsurprisingly, inventive and sophisticated philosophers like
Locke often need to be read philosophically if we are to grasp their
meanings, link their ideas and so on. Just as political theorists and philo-
sophers should think historically, or at least consult historical scholarship,
so too historians should think philosophically, or at least consult philoso-
phical scholarship. Most political theorists and philosophers would now
take seriously contextually driven hypotheses of what Locke meant by
‘property’ or ‘rights’: I think more historians should place greater weight
on philosophically driven hypotheses too.
I wonder if it is revealing that Dunn uses a visual metaphor when arguing
that historians, political theorists and philosophers can all ‘see’ deeper
structural features of Locke if they ‘look closely’ (2005: 442–4). Dunn is
arguing, rightly, that no disciplinary approach need stop us from grasping
Locke’s ideas. But thinking through authors’ ideas, or robust contextual
analysis, obviously takes conscious effort. Even the terms I often use in this
chapter – ‘uncovering’ or ‘recovering’ authorial beliefs and meanings,
following Skinner’s (2002a) terminology – does not capture the hard
reasoning often needed for empirical and systematic reconstruction (exem-
pliﬁed superbly by Dunn himself – e.g. Dunn 1967: 156–82).
I now turn to the aspect of Simmons’s book that is most controversial,
at least for historians: his attempt to improve Locke with a ‘revised
Lockean account’ of such things as the natural right to punish (1992:
121–40, 161–6). Indeed, one of Simmons’s aims is to uncover ‘not
Locke’s own theory . . . but the best version of that theory – a theory
close enough to Locke’s to be considered “Lockean”, but improved by
certain departures from the letter of Locke’s theory’ (1992: 4; emphasis
removed). Most importantly, his Lockean theory of rights drops Locke’s
theological foundations (1992: 3, 10–11, 153–66, 177–204, 343–54).
Simmons is explicit that his adaptations ‘capture some of the spirit’ of
Locke’s moral theory, but are not ‘fully consistent’ with it: ‘God is too
much at the center of Locke’s work for such secular, Kantian arguments
to capture its essence’ (1992: 44, 46).
But what is the ‘best’ version of an author’s theory, or at least, a ‘better’
version? Historians will probably seek answers that the author could have
accepted (e.g. Dunn 1967: 166–8), but need not deny that other answers
are legitimate (e.g. Dunn 1969: 214–16). Political theorists and
philosophers can also opt for answers that the author could have accepted
(e.g. Martinich 2005: 101–4), but may ultimately prefer answers that
many of us could accept today (e.g. Simmons 1992: 10). There will be
some overlap here: most writers, past and present, would want to avoid
Interpreting Texts 257

egregious contradictions and factual or logical errors. But ultimately, his-

torical and contemporary answers may differ markedly (Dunn 1967: 159).
There is no one ‘right’ adaptive reconstruction.
Can we justiﬁably drop authors’ deeply held commitments? Dunn
holds that ‘an extremely high proportion of Locke’s arguments’ have
an ‘intimate dependence’ on his theological commitments (1969: xi).
Simmons replies, in effect, that Dunn’s systematic reconstruction is
overly contextualized and misconstrues Locke’s philosophical posi-
tion: Locke actually over-determines his conclusions, intentionally
and consciously making theological and secular arguments, which
are logically separable (e.g. Simmons 1992: 10–12, 45, 101–2, 254,
354). But if Dunn is right and Simmons is wrong, would secularizing
Locke still be permissible? Similarly, though not identically, should we
drop republicans’ gender assumptions when appropriating republican
liberty for contemporary purposes (e.g. Pettit 1997: 138–40), even
though this would have appalled most historical republicans?
Do such adaptations go too far?
For some people, yes. But two caveats are crucial here. First, the adapta-
tions look less radical once we accept that we cannot understand authors
without altering what they said, that we must sometimes make choices
when authors are inconsistent, and that we often want to impose more
clarity on authors than they themselves managed, as I argued
previously. Second, even leaving aside contemporary applications of his-
torical ideas, it is intellectually interesting and important to see how well
authors’ arguments work, to what extent they were on the right track and
how fundamental their problems are – whether Locke’s ambiguities, gaps,
contradictions and errors can be corrected within his system or not, say.
We should try not to depict such alterations as if they are true to the
author. Unfortunately, this is quite common, as with caricatures of
Hobbes by many international relations scholars (criticized by Malcolm
2002: 432–56). By contrast, Gray is explicitly agnostic about whether his
reconstruction of Mill on happiness is what Mill intended (1996: 70–86).
When discussing Hobbes’s account of representation, similarly, Skinner
infers what Hobbes ‘seems to have had in mind’, and offers the ‘daring’
view that ‘the best statement of Hobbes’s theory is the one that he never
explicitly gave’ (2002b: 190; see also 217, 235). This is not actually too
daring: we consciously or subconsciously improve on what other people
say and write every time we hear or read them. But that quibble aside,
Skinner’s carefulness is absolutely right. Try to be explicit, or at least
implicit, when distinguishing between what authors meant, what
they may have meant, what their account implies and what, in
your view, they should have said.
258 Chapter 12

5 Theoretical and Normative Perspectives

I now turn to a different style of analysis. Again I question traditional
depictions of the methodological issues.
Consider Susan Okin’s feminist perspective, which helps her spot pre-
suppositions and implications in many authors, including those claiming
to be gender-neutral (Okin 1989: 10–13, 44–60, 80–7, 90–7). Or
consider C. B. Macpherson’s Marxist perspective, which leads him to
infer Hobbes’s conscious or unconscious assumptions about ‘possessive
individualism’ and other capitalist traits (1962: 4–5, 26–9, 37–40, 46,
59–68, 78–80, 84–106; see Townshend 1999 for a defence of
Macpherson against common complaints, and Svacek 1976 for whether
Macpherson’s perspective is really Marxist).
Both authors use theoretical and normative perspectives that are often
presented as essentially different ‘schools’ or ‘approaches’, as with Ball’s
comparison of feminist, Marxist, totalitarian, psychoanalytic, Straussian
and postmodernist interpretations (Ball 2004, 19–27; see also Ball 2011).
I will argue, rather, that these analysts do the same thing: they use theories
and norms to develop empirical hypotheses or conceptual distinctions
that help us uncover assumptions and implications.
Seen like this, there are far more perspectives than are usually men-
tioned, even by Ball. For example, Skinner’s republican perspective alerts
him to ideas often neglected in Machiavelli (Skinner 1990: 300–6). John
McCormick’s democratic perspective highlights features in Machiavelli
overlooked by republicans (2011: see especially 3, 7–11, for the critique
of republican interpretations). Hayek’s libertarian/classical liberal per-
spective uncovers more individualism in Burke than other writers saw
(1948: 4–8, 13, 24). Terrell Carver’s gender perspective reveals assump-
tions about men passed over by feminist scholars concentrating on
assumptions about women (2004). David Armitage’s international per-
spective pinpoints oft-overlooked issues in Hobbes and Locke (2013:
62–7, 79–85). Jon Elster’s analytical approach, combining ideas such as
methodological individualism and rational choice theory, pinpoints for-
gotten features of Marx (Elster 1985: see especially 3–48 on Elster’s
analytical tools). Martin Hollis’s game-theoretic approach provides
powerful insights into Hobbes, Hume, Smith, Kant and others (1998).
And so on.
A contribution I ﬁnd especially interesting, and worrying, is Robert
Bernasconi’s race/ethnicity perspective. He uncovers assumptions about
race in Locke, Kant, Hegel, Mill, Nietzsche and others (2003: 14–20;
2010: 500–4, 510–11, 515–16). ‘Western philosophy has been and is still
largely in denial about its racism’, he writes (2003: 14), challenging those
Interpreting Texts 259

of us who have missed these writers’ explicit or implicit racism, or passed

over it in silence in our writing and teaching.
These perspectives, then, can be incredibly powerful. They are one way
we keep seeing new things in old texts. But should we depict different
perspectives as different ‘schools of thought’ or different ‘approaches’?
Throughout this chapter, I have tried to ask what our interpretations
really involve. And in my view, theoretical and normative perspectives
boil down to approaching a text with one or more hypotheses or distinc-
tions, and potentially seeing new things.
I am not disparaging such research: do not underestimate the
insights you can get by applying an empirical hypothesis or
conceptual distinction from an existing theoretical/normative
perspective or elsewhere. In fact, you need not wholly accept
a perspective to apply it. You don’t have to be a republican to ask if
authors uphold freedom as non-domination, a feminist to uncover
authors’ gender presuppositions or a poststructuralist to apply
Foucault’s distinction between a governmentality of politiques and of
économistes (2009: 333–57). You don’t have to be a Marxist to notice
authors’ socio-economic presuppositions. Indeed, it may help not to be
a Marxist, because Marxists might be inclined towards certain conclu-
sions, potentially infecting their interpretations.
Two related dangers thus need attention. The first is that a perspective
leads you to misread a text, read too much into it or overlook relevant
passages – a common problem (Ball 2004: 21–3). Arthur Melzer shows
that most scholars dislike the idea of esoteric writing (i.e. writing that
hides messages between the lines, for careful observers to infer – see
Section 6) and thus overlook ample evidence of esotericism (2014:
especially 13–24, 137–42, 299–317). Yet he himself is so keen to show
Rousseau’s sympathies for esotericism that he overlooks Rousseau’s cri-
tical comments about esotericism (Melzer 2014: 163; Blau 2015b: 163).
Such writers ‘mistake an expectation for a presumption’ (Bevir 1999:
147). Meanwhile, Robin Douglass (2015: 283) argues that Skinner’s
‘preoccupation’ with Hobbes’s republican context leads to misinterpreta-
tions: Skinner’s reading ‘conceals more than it reveals about [Hobbes’s]
battle with republicanism’. If one has a normative axe to grind, one often
chops off key parts of texts. Be attentive to potential theoretical/
normative biases: try to be impartial. If your perspective gives
you an expectation about an author’s influences or motives, it is
only an expectation – a hypothesis – never a certainty.
The second danger is becoming a mouthpiece for flawed ideas. If you
can, test perspectives, don’t just apply them. For example, many
scholars apply Berlin’s flawed distinction between negative and positive
260 Chapter 12

liberty without mentioning its inadequacies, thus overlooking aspects of

liberty noticed by republican scholars (Pettit 1997: 17–50). Showing
a perspective’s shortfalls, and ideally, refining and improving it, magnifies
your work’s value and could expand your audience. For example, Anne
Brunon-Ernst (2012) not only applies Foucault’s ideas of biopolitics to
Bentham, but criticizes and amends Foucault’s ideas in the process.
Both dangers apply to any interpretation: everyone interprets
everything through many lenses (Bevir 1999: 92–3; Brewer and
Lambert 2001). So, try to be aware, if you can, of perspectives
that already influence your readings. In my ongoing Hobbes work,
I tried to chart all of Hobbes’s practical proposals for averting a state of
nature. Yet my mind-set was not attuned to international issues, and after
reading Armitage’s (2013) work on international perspectives, I started
noticing internationally oriented proposals in Hobbes that I had pre-
viously missed. We all have such biases. Even supporting an interpreta-
tion can have this effect: part of you may want it to be right, potentially
infecting your reasoning.

6 Reading Between and Outside the Lines

This chapter has regularly challenged our inherited categories. The same
applies to the idea that ‘reading between the lines’ is an essentially
different approach. Despite claims that some scholars read between the
lines while others take a purely literal approach (Melzer 2014: 112–14,
207, 368), no one takes a purely literal approach, restricting herself to the
actual words: all communication and all textual interpretation involves
reading between the lines, as Section 4 showed. Contextualists agree. For
example, The Prince’s attack on humanist mirror-of-princes handbooks
‘cannot be discovered by attending to Machiavelli’s text, since this is not
a fact contained in the text’ (Skinner 2002a: 143). In effect, when
Machiavelli discusses virtú, we often insert the words ‘unlike the views
of my contemporaries’ into his comments. Similarly, many great histor-
ical authors were religiously unorthodox, but could not say so publicly; we
can sometimes read between the lines and infer what they really thought,
because they cannot conceal all of their views or because they left subtle
signs about their real thoughts (e.g. Schotte 2015: 65–72).
Great care is needed here: without sensible guidelines, we can easily
read too much into texts, especially when asking if authors esoterically
hid messages in their writings for clever readers to spot. Some authors
certainly did this (Patterson 1991). But many esoteric interpretations
overreach themselves, with highly questionable use of evidence (Blau
2012).
Interpreting Texts 261

Careful readers will note that this section has not yet mentioned Leo
Strauss. Those who have read between the lines may already see what
I now state explicitly: we should not equate ‘esoteric’ and ‘Straussian’
interpretation. From the 1940s on, Leo Strauss and his followers made
esoteric interpretations of writers like Plato, Machiavelli, Rousseau and
Nietzsche. There is nothing wrong with esoteric interpretation, but much
wrong with Strauss’s esoteric interpretation – due not to its esotericism,
but to its naive and flawed methodology (Blau 2012; see also Blau 2015b
for some corrections to the original argument).
Unfortunately, our methodological lexicon has clouded the real
problems. We should not say that Strauss has a different hermeneutic
or particular techniques of reading texts. Rather, he has hypotheses
about the particular ways that authors hid messages, and inadequate
tests of these hypotheses (Blau 2015b: 32–40). Fortunately, Melzer’s
wonderful practical advice (2014: 288–99, 323–4) warns Straussians
not to infer esotericism too hastily. Unfortunately, Melzer’s own
handling of claims about esotericism may still encourage such hasti-
ness (Blau 2015a: 162–3).
The excesses of Straussian readings have helped and hindered the
cause of esoteric interpretation. They have helped it by highlighting
a largely forgotten kind of writing and offering evidence of esoteric
techniques. But they have hindered it through methodologically flawed
over-interpretations that have given esoteric interpretation a bad name –
‘Straussian’. Esoteric interpretation, like any empirical interpreta-
tion, is only a hypothesis. And the failure to provide fundamental
principles for testing hypotheses, through a flawed focus on supposedly
separate approaches, is one of the great tragedies of twentieth-century
methodological writings. Readers should not think that my critiques of
Straussians are aimed only at Straussians.

7 Further Core Principles of Good Practice

I have already identiﬁed certain core principles for all textual interpreters,
such as reading texts both contextually and philosophically. I now sum-
marize other core principles.
The underlying ideas are uncertainty and under-determination: no
empirical claims can be known for certain, and the same evidence
can always be read differently. We even disagree about what counts as
evidence. Uncertainty and under-determination are more fundamental
than ‘approaches’ like contextualism or Straussianism: claims about the
relevance of a particular context or the use of a speciﬁc esoteric technique
are only ever hypotheses, and uncertainty and under-determination are
262 Chapter 12

pervasive whenever hypotheses are tested. I expand on these ideas

elsewhere (Blau 2011, 2012, 2015b, 2015c).
Uncertainty has two main implications. First, be careful of
overconfident claims. Try not to talk of ‘proving’ anything, and
where relevant, indicate how confident you are in your interpreta-
tions. There are differences between saying, ‘Machiavelli wrote
The Prince’, which we have no good reason to doubt; ‘Mill does not
seem to have had a single, clear idea of harm in On Liberty’, a plausible
inference; and ‘there is evidence both for Hobbes’s atheism and for his
being a believer, but I find the evidence for the latter stronger than for the
former’, a safe stance given the highly contestable evidence. Do not see
yourself as reporting facts, but as reporting your confidence in
your inferences. The inherent subjectivity of empirical research means
that however stylistically ugly you find this, you may need to put the
focus on you, not on the text/author, e.g. ‘I suspect that Cicero meant
P’, ‘Cicero could have meant P or Q, but P seems likelier’ and so on (Blau
2011: 362–8).
Second, and more important, uncertainty often requires us to test
empirical claims. Under-determination kicks in here, so your simplest
and best test of a claim is to see if it fits the evidence better than
plausible alternatives. Simplifying somewhat, you should consider
what fits your interpretation and what does not, and also what fits alter-
native interpretations and what does not. Interpretation is comparative
(Blau 2015c: 1184).
The secondary literature is usually a good source of alternative inter-
pretations. Addressing other scholarship is not something we do as an
offering to the footnote gods: we need to see if other scholars interpret
ideas differently or have spotted things we have missed. Critically compar-
ing interpretations is simultaneously defence and offence, supporting one’s
account by showing that it works better than the alternatives (if space
permits). But your initial expectations may not endure. Don’t become
attached to an interpretation because it’s yours; become attached
to an interpretation because you think it’s better than the
alternatives. The two won’t always go together, unless you are
staggeringly clever or astonishingly lucky.
A powerful test is to triangulate evidence by seeing if textual,
contextual, philosophical and motivational evidence imply the
same conclusions. Textual and contextual evidence have been amply
discussed earlier. Philosophical evidence refers to such things as the
implications of arguments or the consistency of two ideas. Might the
implications of Rousseau’s comments of liberty imply his motivations –
are his definitions implicitly undermining the positions of other authors?
Interpreting Texts 263

Do the implications of his comments on civil and moral freedom help us

understand which, if either, applies to ‘forced to be free’?
Motivational evidence means inferences about authors’ motivations,
which can provide further evidence in our investigations. For example,
book 1 of the Republic uses Socrates’s style of argument, ‘elenchus’, but
books 2–10 use Plato’s own approach, ‘dialectic’: perhaps Plato was
subtly showing Socrates’s limitations (Reeve 1988: 3–24). There are
other possibilities too, e.g. that the text was written at different times,
and we read the Republic differently depending on our stance here. Note
that we cannot see motivations: we only infer them from textual, con-
textual and/or philosophical evidence.
Ideally we want as much evidence as possible. Imagine that textual and
philosophical analysis gives you a possible solution to Mill’s confusing
comments about offence (see Section 4). Think about seeking contextual
evidence: might a controversy at the time explain Mill’s comments, e.g.
parliamentary debates about public drug-taking or masturbation? Finding
such a link would strengthen your inference about what Mill had in mind.
Not finding one would not weaken your position, as Mill might plausibly
have discussed such an example anyway. Or perhaps you are a historian
who has uncovered such a contextual event. If it fits philosophically with
the relevant passages in Mill, your position is strengthened; if not, you
might rethink. (For more on such uses of ‘observable implications’ to test
ideas, see Blau 2012: 152; 2015b: 35–6. This draws on the hypothesis-
testing approach of Van Evera 1997: 31–2.)
We do not always have the time or energy for this. Being corrected on
these grounds is to be expected. My fear, though, is that our inherited
disciplinary and methodological categories often constrain us: seeing
ourselves as contextualists, philosophers or Straussians, say, can deflect
us from important evidence. True, we all have certain skills, and while this
partly reflects disciplinary training, some people just are better at concrete
historical research or abstract philosophizing. Even then, we can read the
expert literature from another field, consult experts or find a co-author.
Ultimately, though, if someone neglects relevant evidence, others can
supply it and test the argument: triangulation can be communal.
Our evidence should be visible. Some scholars follow the bad academic
convention of giving page numbers only for quotations, but not ideas.
This can facilitate caricatures. Yet we should not robotically give page
numbers alone where there are many different editions and translations
(e.g. for Rousseau’s Social Contract). Further details are needed here.
The principle is thus: make clear what your evidence is so that others
can easily follow it up and test your claims.
264 Chapter 12

Another aspect of clarity involves interpretation that is both empirical

and conceptual. Strictly speaking, this is always the case (Bevir 1999: 98;
Skinner 2002a: 16, 45). But sometimes conceptualization is especially
important, as when we ask how ‘modern’ Machiavelli was or how ‘liberal’
Rousseau was. In such cases, conceptual clarity is vital. Harry Lesser’s
(1979) short article on Plato’s feminism does not say what he means by
‘feminism’, although we could probably make some inferences from his
account. Virginia Sapiro, by contrast, is much conceptually clearer when
discussing Wollstonecraft’s feminism (1992: 258–9). When analysis is
partly conceptual and partly empirical, try to define key terms (see
Olsthoorn’s chapter in this volume, Section 3).
The terms in quotation marks in the preceding paragraph are anachro-
nistic. Care is needed here, as Section 3 noted. For example, Michael
Losonsky (2001: 53–4) talks of ‘deliberative reason’, ‘passionate reason-
ing’ and ‘reasoned deliberation’ in Hobbes. These terms misconstrue
Hobbes’s position (Blau 2016: 209–10). By contrast, David Wootton
asks how democratic the Levellers were by applying careful conceptual
distinctions to rigorous analysis of their writings (1992: 71–80).
Skinner finds such anachronisms ‘pointless’ (2002c: 58), but this might
reflect his particular view of meaning and understanding – a purely
empirical view (‘the Levellers meant X’) in contrast to a philosophical
view (‘their comments fit modern democratic criteria, which means that
we can call them democrats’). Both types of meaning matter: the first lets
us understand what the Levellers meant, the second lets us understand
their originality and importance. Skinner’s criticism of anachronisms
seems to reflect his dependency on a single notion of meaning and under-
standing. This chapter has sought to liberate us from the dominant
approach.

8 Conclusion
This chapter’s key guidance can be summarized as follows: Read widely
and carefully. Think contextually and philosophically. Embrace
uncertainty. See both sides. Think against yourself. Question
evidence and interpretations. Test. Retest. Be open. And be open-
minded. Such how-to guidance – sometimes obvious, sometimes not – is
largely absent from the existing methodological literature, because of the
tendency to discuss different approaches and schools of thought without
also emphasizing core principles of good interpretation. I do not mean to
deride our methodological literature, and would especially encourage
novice interpreters to pore over Skinner’s methodological writings – but
also his substantive interpretations. My ﬁnal suggestion is thus: see
Interpreting Texts 265

methodology as something that you do not learn merely from

methodological writings. Reﬂect also on what is good and less
good in actual interpretations, and infer principles of good
practice from that.

Acknowledgements
For comments and criticisms on earlier versions of this chapter, I thank
Seebal Aboudounya, Richard Bourke, Emillie De Keulenaar, Robin
Douglass, John-Erik Hansson, Eva Hausteiner, Johan Olsthoorn,
Joanne Paul, Mustafa Rehman, Paul Sagar, Dave Schmidtz, Max
Skjönsberg, Bertie Vidgen and Sarah Wilford.

References
Allen, J. W., 1928. A History of Political Thought in the Sixteenth Century. London:
Methuen.
Armitage, David, 2013. Foundations of Modern International Thought. Cambridge
University Press.
Ball, Terence, 1992. ‘Introduction’, in James Mill, Political Writings, ed.
Terence Ball. Cambridge University Press, ix–xxviii.
Ball, Terence, 2004. ‘History and the interpretation of texts’, in Gerald Gaus and
Chandran Kukathas, eds., Handbook of Political Theory. London: Sage
Publications, 18–30.
Ball, Terence, 2011. ‘The value of the history of political philosophy’, in
George Klosko, ed., The Oxford Handbook of the History of Political Philosophy.
Oxford University Press, 47–59.
Ball, Terence, James Farr and Russell Hanson, eds., 1989. Political Innovation and
Conceptual Change. Cambridge University Press.
Berlin, Isaiah. 1969. Four Essays on Liberty. Oxford University Press.
Bernasconi, Robert, 2003. ‘Will the real Kant please stand up: the challenge of
Enlightenment racism to the study of the history of philosophy’, Radical
Philosophy 117, 13–22.
Bernasconi, Robert, 2010. ‘The philosophy of race in the nineteenth century’, in
Dean Moyar, ed., The Routledge Companion to Nineteenth Century Philosophy.
London: Routledge. 498–521.
Bevir, Mark, 1999. The Logic of the History of Ideas. Cambridge University Press.
Blau, Adrian, 2009. ‘Hobbes on corruption’, History of Political Thought 30,
596–616.
Blau, Adrian, 2011. ‘Uncertainty and the history of ideas’, History and Theory 50,
358–72.
Blau, Adrian, 2012. ‘Anti-Strauss’, The Journal of Politics 74, 142–55.
Blau, Adrian, 2015a. ‘Philosophy between the lines, or through dubious signs?’,
Perspectives on Political Science 44, 162–5.
266 Chapter 12

Blau, Adrian, 2015b. ‘The irrelevance of (Straussian) hermeneutics’, in

Winfried Schröder, ed., Reading Between the Lines: Leo Strauss and the History
of Early Modern Philosophy. Berlin: De Gruyter, 29–55.
Blau, Adrian, 2015c. ‘History of political thought as detective-work’, History of
European Ideas 41, 1178–94.
Blau, Adrian, 2016. ‘Reason, deliberation and the passions’, in A. P. Martinich and
Kinch Hoekstra, eds., The Oxford Handbook of Hobbes. Oxford University Press,
195–220.
Boucher, David, 1985. Texts in Context: Revisionist Methods for Studying the History
of Ideas. Dordrecht: Kluwer Academic Publishers.
Brewer, William and Bruce Lambert, 2001. ‘The theory-ladenness of observation
and the theory-ladenness of the rest of the scientiﬁc process’, Philosophy of Science
68 supplement, 176–86.
Brunon-Ernst, Anne, 2012. Utilitarian Biopolitics: Bentham, Foucault and Modern
Power. London: Pickering & Chatto.
Buchan, Bruce and Lisa Hill, 2014. An Intellectual History of Political Corruption.
Basingstoke: Palgrave Macmillan.
Carver, Terrell, 2004. Men in Political Theory. Manchester University Press.
Cohen, Marshall, 1960. ‘Berlin and the liberal tradition’, The Philosophical
Quarterly 10, 216–27.
Douglass, Robin, 2015. ‘Thomas Hobbes’s changing account of liberty and chal-
lenge to republicanism’, History of Political Thought 36, 281–309.
Dunn, John, 1967. ‘Consent in the political theory of John Locke’, The Historical
Journal 10, 153–82.
Dunn, John, 1968. ‘The identity of the history of ideas’, Philosophy 43, 85–104.
Dunn, John, 1969. The Political Thought of John Locke: An Historical Account of
The Argument of the ‘Two Treatises of Government’. Cambridge University Press.
Dunn, John, 1996. The History of Political Theory and Other Essays. Cambridge
University Press.
Dunn, John, 2005. ‘What history can show: Jeremy Waldron’s reading of Locke’s
Christian politics’, The Review of Politics 67, 433–50.
Elster, Jon, 1985. Making Sense of Marx. Cambridge University Press.
Euben, Peter, 1989. ‘Corruption’, in Terence Ball, James Farr and
Russell Hanson, eds., Political Innovation and Conceptual Change. Cambridge
University Press, 220–46.
Forrester, Katrina, 2014. ‘Citizenship, war, and the origins of international ethics
in American political philosophy, 1960–1975’, The Historical Journal 57,
773–801.
Foucault, Michel, 2009. Security, Territory, Population: Lectures at the Collège de
France 1977–1978, tr. Graham Burchell. Basingstoke: Palgrave Macmillan.
Gray, John, 1995. Berlin. London: Fontana Press.
Gray, John, 1996. Mill On Liberty: A Defence. 2nd edition. London: Routledge.
Green, Jeffrey, 2015. ‘Political theory as both philosophy and history: a defense
against methodological militancy’, Annual Review of Political Science 18, 425–41.
Interpreting Texts 267

Griswold, Charles, 2016. ‘Genealogical narrative and self-knowledge in

Rousseau’s Discourse on the Origin and the Foundations of Inequality among
Men’, History of European Ideas 42, 276–301.
Hayek, Friedrich, 1948. Individualism and Economic Order. University of Chicago
Press.
Hobbes, Thomas, 1991. Leviathan, ed. Richard Tuck. Cambridge University
Press.
Hobbes, Thomas, 1998. On the Citizen, ed. Richard Tuck and
Michael Silverthorne. Cambridge University Press.
Hollis, Martin, 1998. Trust Within Reason. Cambridge University Press.
Hutton, Sarah, 2014. ‘Intellectual history and the history of philosophy’, History of
European Ideas 40, 925–37.
Kelly, Paul, 2006. ‘Political theory – the state of the art’, Politics 26, 47–53.
Kelly, Paul, 2007. Locke’s Second Treatise of Government: A Reader’s Guide.
London: Continuum.
Lane, Melissa, 2012. ‘Doing our own thinking for ourselves: on Quentin Skinner’s
genealogical turn’, Journal of the History of Ideas 73, 71–82.
Laslett, Peter, 1988. ‘Introduction’, in John Locke, Two Treatises of Government,
ed. Peter Laslett. Cambridge University Press, 3–126.
Lesser, Harry, 1979. ‘Plato’s Feminism’, Philosophy 54, 113–17.
Losonsky, Michael, 2001. Enlightenment and Action from Descartes to Kant:
Passionate Thought. Cambridge University Press.
Macpherson, C. B., 1962. The Political Theory of Possessive Individualism: Hobbes to
Locke. Oxford University Press.
Malcolm, Noel, 2002. Aspects of Hobbes. Oxford: Oxford University Press.
Martinich, A. P., 2005. Hobbes. London: Routledge.
Mason, John Hope. 1995. ‘Forced to be free’, in Robert Wokler, ed., Rousseau and
Liberty. Manchester University Press, 121–38.
McCormick, John, 2011. Machiavellian Democracy. Cambridge University Press.
Melzer, Arthur, 2014. Philosophy Between the Lines: The Lost History of Esoteric
Writing. University of Chicago Press.
Mill, J. S., 1989. On Liberty and Other Writings, ed. Stefan Collini. Cambridge
University Press.
Ober, Josiah, 2008. ‘The original meaning of “democracy”: capacity to do things,
not majority rule’, Constellations 15, 3–9.
Okin, Susan Moller, 1989. Justice, Gender, and the Family. New York: Basic Books.
Olsen, Niklas, 2012. History in the Plural: An Introduction to the Work of Reinhart
Koselleck. Oxford: Bergahn Books.
Patterson, Annabel, 1991. Censorship and Interpretation: The Conditions of Writing
and Reading in Early Modern England. 2nd edition. Madison: University of
Wisconsin Press.
Pettit, Philip, 1997. Republicanism: A Theory of Freedom and Government. Oxford
University Press.
Pitkin, Hanna, 1967. The Concept of Representation. Berkeley: University of
California Press.
268 Chapter 12

Pocock, J. G. A., 1971. Politics, Language, and Time: Essays on Political Thought and
History. New York: Atheneum.
Pocock, J. G. A., 2006. ‘Foundations and moments’, in Annabel Brett and
James Tully, eds., Rethinking the Foundations of Modern Political Thought.
Cambridge University Press, 37–49.
Prado, C. G., 2006. Searle and Foucault on Truth. Cambridge University Press.
Rawls, John, 1971. A Theory of Justice. Cambridge, MA: Harvard University Press.
Reeve, C. D. C., 1988. Philosopher-Kings: The Argument of Plato’s Republic.
Princeton, NJ: Princeton University Press.
Richter, Melvin, 1995. The History of Political and Social Concepts: A Critical
Introduction. Oxford University Press.
Richter, William, 2009. ‘Introduction: the study of political thought’, in
William Richter, ed., Approaches to Political Thought. Lanham, MD: Rowman
& Littleﬁeld, 1–12.
Rorty, Richard, 1984. ‘The historiography of philosophy: four genres’, in
Richard Rorty, J. B. Schneewind and Quentin Skinner, eds., Philosophy in
History: Essays on the Historiography of Philosophy. Cambridge University Press,
49–75.
Rosenblatt, Helena, 1997. Rousseau and Geneva: From ‘the First Discourse to
The Social Contract, 1749–1762. Cambridge University Press.
Rousseau, Jean-Jacques, 1979. Emile or on Education, ed. Allan Bloom. New York:
Basic Books.
Rousseau, Jean-Jacques, 1997a. ‘ The Social Contract,’ in The Social Contract and
Other Later Political Writings, ed. Victor Gourevitch. Cambridge University
Press.
Rousseau, Jean-Jacques, 1997b. ‘Considerations on the Government of Poland’,
in The Social Contract and Other Later Political Writings, ed. Victor Gourevitch.
Cambridge University Press.
Sapiro, Virginia, 1992. A Vindication of Political Virtue: The Political Theory of Mary
Wollstonecraft. University of Chicago Press.
Schotte, Dietrich, 2015. ‘The virtues and vices of Leo Strauss, historian.
A reassessment of Straussian hermeneutics’, in Winfried Schröder, ed.,
Reading Between the Lines: Leo Strauss and the History of Early Modern
Philosophy. Berlin: De Gruyter, 57–76.
Schulz, Daniel and Alexander Weiss, 2010. ‘Introduction: approaches in the
history of political thought’, European Political Science 9, 283–90.
Searle, John, 1978. ‘Literal meaning’, Erkenntnis 13, 207–24.
Sidgwick, Henry, 1981. The Methods of Ethics. Seventh edition. Indianapolis, IN:
Hackett.
Simmons, A. John, 1992. The Lockean Theory of Rights. Princeton, NJ: Princeton
University Press.
Skinner, Quentin, 1964. ‘Hobbes’s “Leviathan”’, The Historical Journal 7, 321–33.
Skinner, Quentin, 1981. Machiavelli. Oxford University Press.
Skinner, Quentin, 1989. ‘The state’, in Terence Ball, James Farr and
Russell Hanson, eds., Political Innovation and Conceptual Change. Cambridge
University Press, 90–131.
Interpreting Texts 269

Skinner, Quentin, 1990. ‘The republican ideal of political liberty’, in Gisela Bock,
Quentin Skinner and Maurizio Viroli, eds., Machiavelli and Republicanism.
Cambridge University Press, 293–309.
Skinner, Quentin, 1996. Reason and Rhetoric in the Philosophy of Hobbes. Cambridge
University Press.
Skinner, Quentin, 2002a. Visions of Politics. Volume I: Regarding Method.
Cambridge University Press.
Skinner, Quentin, 2002b. Visions of Politics. Volume III: Hobbes and Civil Science.
Cambridge University Press.
Skinner, Quentin, 2002c. ‘Interview with Quentin Skinner’, Finnish Yearbook of
Political Thought 6, 32–63.
Skinner, Quentin, 2003. ‘States and the freedom of citizens’, in Quentin Skinner
and Bo Stråth, eds., States and Citizens: History, Theory, Prospects. Cambridge
University Press, 11–27.
Skinner, Quentin, 2008. Hobbes and Republican Liberty. Cambridge University
Press.
Skinner, Quentin, 2009. ‘A genealogy of the modern state’, Proceedings of the British
Academy 162, 325–70.
Skinner, Quentin, 2012. ‘On politics and history: a discussion with Quentin
Skinner’, Journal of Intellectual History and Political Thought 1, 7–31.
Sreenivasan, Gopal, 2000. ‘What is the General Will?’, The Philosophical Review
109, 545–81.
Stuurman, Siep, 2000. ‘On intellectual innovation and the methodology of the
history of ideas’, Rethinking History 4, 311–19.
Svacek, Victor, 1976. ‘The elusive Marxism of C. B. Macpherson’, Canadian
Journal of Political Science 9, 395–422.
Townshend, Jules, 1999. ‘Hobbes as possessive individualist: interrogating the
C. B. Macpherson thesis’, Hobbes Studies 12, 52–72.
Tuck, Richard, 1993. Philosophy and Government 1572–1651. Cambridge
University Press.
Tully, James, 1995. ‘Property, self-government and consent’, Canadian Journal of
Political Science 28, 105–32.
Van Evera, Stephen, 1997. Guide to Methods for Students of Political Science. Ithaca,
NY: Cornell University Press.
Waldron, Jeremy, 2002. God, Locke, and Equality. Christian Foundations of John
Locke’s Political Thought. Cambridge University Press.
Whatmore, Richard, 2006. ‘Intellectual history and the history of political
thought’, in Richard Whatmore and Brian Young, eds., Palgrave Advances in
Intellectual History. Basingstoke: Palgrave Macmillan, 109–29.
Wolff, Jonathan, 2013. ‘Analytic political philosophy’, in Michael Beaney, ed.,
The Oxford Handbook of the History of Analytic Philosophy. Oxford University
Press, 795–822.
Wootton, David, 1992. ‘The Levellers’, in John Dunn, ed., Democracy:
The Unﬁnished Journey. Oxford University Press, 71–89.
Zagorin, Perez, 1993. ‘Hobbes’s Early Philosophical Development’, Journal of the
History of Ideas 54:3, 505–18.
13 Comparative Political Thought

Brooke Ackerly and Rochana Bajpai

1 Introduction
Two features characterize the emerging field of comparative political
thought (CPT). The first is an interest in bringing Islamic, Indian,
Chinese, African and Latin American ethical-political thought to the
attention of Western audiences. The second is an interest in combining
the methods of political theorists with those of historians, political
scientists and anthropologists to enhance real-world relevance of political
theory. This chapter provides an overview of normative-analytic, histor-
ical, interpretive and critical approaches to CPT with examples of var-
iants within each category. It argues for a methodology-based definition
of CPT, and for a non-hierarchical pluralism in which multiple vocations
of political theory prevail. Recognizing the complexity of questions raised
by the topic of methods in political theory generally, our aims in this
chapter are first to provide an introduction to CPT, second to provide
a schematic of the four methodological strands within CPT and finally to
highlight the range of methods useful for CPT research, and, where these
have something in common with other practices of political theory, to
elaborate on how they are used in CPT. To underline, this is an argument
about the variety of methodologies and methods in CPT scholarship and
an appreciation of the value of this methodological pluralism for the field
overall.
Political theorists have been comparing across time, authors and con-
texts since the field began. In the twenty-first century, given the contribu-
tions of post-colonial, feminist and queer scholars, we would expect most
political theorists to attempt to be self-aware regarding the parochialism
and historical elitism of some political theory. However, today, the phrase
‘comparative political thought’ indicates to an Anglophone audience
a way to teach and practise political theory that is less Western-centric.
Obviously, non-Western-centric approaches to political thought have
long featured in teaching and writing in Asia, Africa, the Middle East
and Latin America. Today, the phrase ‘comparative political thought’ is

270
Comparative Political Thought 271

used to characterize political theory within the Anglo-American academy

that uses methods that challenge Euro-American parochialism.
The difference between contemporary CPT and cross-cultural, cross-
language and cross-time comparisons of previous generations may be its
self-reflective attempts to define the field and methods of CPT.
Early progenitors of CPT sought to emphasize cultural differences and
incomparability. For example, British colonizers engaged with political
elites in South Asia and China to interpret local political theory and
cultural practice in ways that facilitated the political economy of colonial
rule. The post–World War II project of identifying the universal human
rights that provided the anchor commitments of the United Nations,
however, sought to identify universal values that transcended cultural
differences. In the post–Cold War era, the breakup of the former
Yugoslavia and multinational communist regimes, the transformations
in the political economies of East Asia, the ongoing tensions in the Middle
East and the influence of Islamist political thought has led to the demand
for political theory that does not serve imperial ends and is open to
universal dimensions of humanity and yet treats differences as strengths.
Contemporary CPT reflects the incorporation within mainstream poli-
tical theory of the influences of anti-colonial, anti-racist and feminist
political theory (see also Gordon 2014).1 The groundwork for the self-
reflective turn in political theory embodied in CPT was laid by their
criticisms of imperial and patriarchal ideologies. In particular, post-
colonial and feminist criticisms of hierarchal notions of difference on
the one hand, and false universalisms on the other, represented one of
the profound contributions to political theory. CPT reflects a broadening
of the range of intellectual resources being brought to bear on these
critiques beyond the Euro-American core of political theory.
It represents the motivation to develop theories and methods grounded
in the historical traditions and contemporary practices of non-Western
societies that transcend the frameworks of area studies, and at the same
time, are not dominated by liberal theories.
How do we determine which intellectual traditions should inform
contemporary comparative reflections? Should we dialogue between or
among such traditions? Do we need to study multiple languages in order
to participate in such dialogues? Or do we need language fluency along
with deep ethnographic and historical understanding of a particular
unknown conceptual terrain in order to shed light on new problems?
1
Compare this account of the history of the field with those of Freeden (2007), Godrej
(2011) and Von Vacano (2015). On the parallels between Orientalist scholarship, and
comparative political theory in terms of seeking to articulate the value of non-Western
scholarship for Western knowledge, see Thomas (2010).
272 Chapter 13

Or, does CPT require engagement with contemporary scholars who have
training, experience and imaginations developed in multiple traditions?
As all of these questions suggest, many scholars today feel the need to
broaden the tool-kit of political theory to include not only new languages,
but also new ways of thinking about the historical epistemology of con-
cepts. These are questions about methods.
CPT scholars use a range of methods and answer these questions
differently. Methods choices depend on the researcher’s ontology,
epistemology and methodology. Ontology is the worldview that deter-
mines what the theorist finds to be an interesting question; epistemology
is the theoretical distinction between fact and opinion and is behind
a theorist’s choice of methodology; methodology is the theory of how
research should proceed. Methods are the specific techniques or tools for
carrying out the research (Ackerly and True 2010: 9–10; 2012). Certain
methods – for example the close reading of historical texts – are used by all
methodologies of CPT. However, these methods are used differently
depending on the specific research question and overarching methodol-
ogy of the project.
In order to discuss the methodologies and methods of CPT, we have to
engage, but not embrace, two suspect premises: that the field of compara-
tive political thought is a subdiscipline with distinct methods and that
delimiting those methods strengthens the field. Perhaps because
comparative political thought is a relatively new field in political theory,
most participants in the field have spent significant time being explicit
about and defending their methods. Michael Freeden refers to ‘the
absence of considered approaches, and their attendant methodologies’
of CPT as ‘one of the great lacunae in the study of political thinking’
(2007: 1). For many political theorists, explaining and defending our
choice of and use of chosen methods is part of justifying our arguments
(Herzog 1985; Gaus 1990; Forst 2001). We believe that a discussion of
methodologies is useful in order to clarify the choices available and to
defend the intellectual case for all of the methods in the field. What the
field needs now is a pluralist account of those methodologies so that
rather than debating which methods and approaches the field should
privilege, we can build an inclusive field whose methodological attentions
focus on the importance of (1) choosing the approach or approaches
appropriate to the particular question at hand and (2) doing that
approach well.
In this chapter we discuss a range of normative, historical, interpretive
and critical approaches to comparative political thought. Not all of those
who participate in the field use all of these approaches, and some CPT
scholars would not recognize certain approaches as being appropriately
Comparative Political Thought 273

understood as part of the ﬁeld. However, we have taken a pluralist under-

standing of the ﬁeld and thus review a broad range and their methodolo-
gical implications. To prescribe how comparative political thought
should be done from the standpoint of one particular methodology or
method is to truncate prematurely the scope of this emerging ﬁeld. In this
sense, our view of CPT is critical of non-pluralist approaches; however,
we treat these approaches, like the others, as offering resources for study-
ing certain questions.

2 What Is Comparative Political Thought?

The defining aim of comparative political thought has been to foster
greater engagement with non-Western thought within mainstream poli-
tical theory. The reasons for studying political thought in the Middle
East, Asia, Africa and Latin America of course differ among CPT scho-
lars. A secondary aim has been to enhance the lived-world relevance of
political theory, in many cases through combining the methods of poli-
tical theorists with those of political scientists, historians, anthropologists
and other empirical disciplines. In their early work in CPT, Fred
Dallmayr and Roxanne Euben engaged with both strands. For
Dallmayr, political theorists were late in responding to the issues raised
by Islamic extremism and globalization, because they neglected real-
world political processes in their preoccupation with abstract theory-
building (2004: 249). Euben noted that the boundaries between political
theory and comparative politics were arbitrary (1999: 159).
The first aim raises a challenging set of questions: is a focus on non-
Western thought sufficient to characterize a study as ‘comparative political
thought’, or does it need to fulfil additional criteria to qualify as compara-
tive inquiry, as some have argued (Freeden 2007; March 2009)? There
has been a long history of centralizing the Western canon, and also of the
post-colonial critique of Eurocentrism within political theory. Does
bringing texts from Asia, Africa and Latin America into the canon then
suffice for comparative inquiry in political thought? In this vein of reflec-
tion, the use of the label ‘comparative political thought’ simply as an
account of a thinker’s geographical location, independent of whether
comparison in a substantive sense is engaged, appears to ghettoize non-
Western political thought (March 2009).
The second aim raises another set of questions for CPT. If the
primary definition of comparative political thought is methodological,
in terms of combining methods of political theory and comparative
politics for instance, is the study of non-Western political thought
necessary for the definition of the field? In other words, does the
274 Chapter 13

signiﬁcance of the method of comparison mean that we should drop

the initial emphasis on non-Western texts as a defining feature of
comparative political theory, while still recognizing that it has great
value (just as some comparisons in political science are more instruc-
tive than others)? In this vein, studies that compare political thought
across Western countries, for instance, would qualify as exercises in
comparative political thought, as is the case in sub-disciplines like
comparative literature or comparative religion (Freeden 2007; Goto-
Jones 2011), as would single-country comparisons of different cases
(Bajpai 2011). In interpretive and critical approaches discussed later,
the initial aim of CPT of engaging with marginalized bodies of poli-
tical thought takes the form of a focus less on non-Western sources
than on non-elite texts and action, usually the subject of comparative
politics, as the ‘texts’ of political theory.
To what extent should we think of both regional and comparative
methods approaches to CPT as part of the same enterprise? We argue
that these are both essential parts of the field. To deny the first is to
divorce CPT from its animating impulse that remains relevant, that of
challenging the marginalization of non-Western forms of knowledge pro-
duction and transmission (Godrej 2011: 14) and potentially transforming
the repertoire of concepts and methods in Western political theory (Jenco
2007). To deny the second is to ignore the pitfalls of a disciplinary
division of labour between political theory and political science (e.g.
Freeden 2007), as well as the politics of disciplinary boundary policing
and crossing, that have always been a part of political theory as well (e.g.
Wolin 1969).
The methodological meaning of the ‘comparative’ in CPT – the joining
up the methods of political theorists and comparative analysts – is equally
central to intellectual engagement with marginalized bodies of thought.
In our view, the comparative enterprise seeks to challenge dominant
ideologies and epistemologies and therefore cannot rely only on tradi-
tional textual resources. As such, it becomes part of the role of CPT to
reflect on methodologies in political theory and the epistemological and
ontological power dynamics these exhibit.
Within CPT, political theory has been construed broadly, to include
analytical political philosophy (sometimes called ‘normative’ theory),
history of political thought, critical theory and interpretation, the study
of political ideologies and discourse analysis (Freeden 2007;
March 2009). Constructing normative arguments about what it is right
to pursue, advancing understanding of thinkers, texts and traditions, as
well as enhancing explanations of political phenomena, all count as
Comparative Political Thought 275

political theory inasmuch as they focus on concepts and arguments that

are at the heart of political theory.
Like political theory, comparison too is construed widely in compara-
tive political thought, across differences of space as well as time. Most
broadly, political theory has been comparative since the time of Aristotle;
indeed, some have argued that political theory is inherently comparative
(Euben 1999; Freeden and Vincent 2013). Most narrowly, CPT might
compare cultural difference between Western and non-Western thought
(March 2009). However, as noted earlier, whether comparison across
a Western/non-Western difference is a necessary criterion for compara-
tive political thought is a matter of continuing debate in the ﬁeld.
CPT destabilizes familiar references points within political theory.
CPT questions can be disruptive and ‘decolonizing’ (e.g. Chan 2009;
Mills 2015), seeking to displace the hegemony of Western categories and
methods (Jenco 2007; Godrej 2011). They can also be transformative and
constructive (e.g. Jenco 2011; Kim 2014). And CPT can be conversa-
tional or discursive, seeking to exchange insights and build new insights
(e.g. Dallmayr 1999; Angle 2002; Ackerly 2005; Bajpai 2011). Some
CPT scholars, trained in Western thought, have no training in the
languages and political thought traditions outside of the West and have
been working outside of their comfort zone for intellectual and pedago-
gical reasons. There are others – often academics from non-Western
traditions – for whom academic training in political thought has always
been cross-contextual. And there are others who have trained in one
tradition but later develop research interests that require another tradi-
tion and extensive language training. Hence, the practice of CPT reﬂects
a range of methodologies.

3 How Do We Do CPT? Four Methodological Approaches

CPT is like a family tree, with participants utilizing a broad range of
methodologies drawing on a complex root system constructed out of
a range of political and intellectual struggles. As the field has developed,
sometimes rigorous defence of the appropriateness of one’s methodology
has meant privileging a particular approach. In contrast, our aim is not so
much to argue for the superiority of a particular approach, but to highlight
the diversity of methodologies and purposes in the field.
We distinguish four principal modes of doing CPT for heuristic
purposes: normative-analytic, historical, interpretive and critical. Our
intention is not to reproduce conventional disciplinary categories, but to
show how these have been configured and challenged in CPT scholar-
ship. Within each methodology, there are of course multiple practices.
276 Chapter 13

In our discussion, we nuance our categorization with speciﬁc examples.

Individual theorists often work in more than one mode: theorists may
emphasize one methodology in one work and another approach in
another work, or, indeed use multiple methodologies in a single study.
Thus, our classification does not seek to label scholars, but rather to
identify the diversity of methodologies that have constituted the field of
CPT, and to suggest exemplary instances of each.
Distinguishing between these approaches is important for at least two
reasons. First, it gives those interested in turning to CPT a sense of how to
draw on their own individual strengths in political theory to become more
comparative in their scholarship. Second, in the field of political theory,
critical and interpretive social science approaches are often mis-
recognized as normative and historical approaches, respectively.
We argue that these are qualitatively distinct in their epistemological
and methodological implications. Interpretive social scientific and critical
methodologies act in part as a check on the epistemological power of
normative-analytic and historical methodologies in the field of political
theory.

3.1 Normative-Analytic CPT

The first category is of CPT as a form of or as an aid to building
normative-analytic theory. One strand among normative-analytic
approaches is a response to the criticism that political philosophers have
been insufficiently attentive to the empirical conditions needed for their
recommendations to be translated into reality (see also the chapters by
Jubb and by Schmidtz in this volume). CPT in this vein attempts to
construct normative arguments that are sensitive to the gap between
‘ideal principles and social reality’ and seek to ‘find a way to connect
facts and norms, practical reforms and substantive ideals’ (Laborde 2008:
13) (cf. Chan 2009: 5). Among scholars of non-Western political
thought, normative-analytic methods are to be found in studies that
compare familiar Western principles such as democracy, rule of law and
human rights with the intellectual resources of non-Western traditions
such as Confucianism (e.g. Ackerly 2005), or use Western terms of
political philosophy – such as ‘ideal’ and ‘non-ideal’ theory – when study-
ing non-Western traditions (see Chan 2009: Introduction; Kim 2014:
chapter 1).
By way of illustration, Stephen Angle (2005) elaborates and juxtaposes
the normative resources for democratic centralism in Rawls’s notion of
a decent society on the one hand, and in Chinese thought, on the other. His
motivation is not only to find ‘grounds for mutual respect’ (Angle 2005:
Comparative Political Thought 277

540), but also to further the cause of practical reforms in contemporary

China. Importantly, although the desired direction is of democratizing
reform, Western liberal democracy is not posited as the desired end-state
in Angle’s account. Instead, the comparative reconstruction of the
Rawlsian notion of a decent society and of Chinese ideas of democratic
centralism in conversation with each other allows for the elaboration of
alternative ideals that are distinct from traditional liberal democracy (and
also from Chinese practice), and crucially, that can be accepted as legit-
imate by those engaged in Chinese politics. In a similar vein but with
different conclusions, Daniel Bell (2000) offers controversial proposals
for a democratic regime based on traditional Confucianism of rule by ‘a
capable and public-spirited “Confucian” intellectual elite’. Like Angle,
Bell affirms ‘the importance of local knowledge of cultural traditions’ not
only ‘from the standpoint of their efficacy, but also from the point of view of
revising political ideals themselves’ (Bell 2000: 19, 14).
In these writings, both theorists draw on empirical knowledge of non-
Western societies and institutions and seek to construct normative
arguments of real-world relevance in an analytic mode. Engagement
with non-Western traditions serves in both cases to enlarge the repertoire
of desirable ideals. Other work by these authors demonstrate a more
explicitly interpretive and even critical CPT methodology – see, for
instance, Bell (2008), Angle (2012) and Kim (2014).
The connection between ideal and non-ideal theory is common among
normative approaches (see also Tan 2003), but it is not the only framing
of normative-analytic CPT. Andrew March in some of his writings offers
a combative statement in support of certain normative-analytic
approaches to non-Western thought. March argues that the ultimate
aim of CPT should be to evaluate moral conflicts between distinct tradi-
tions of thought: ‘comparative political thought derives its greatest sanc-
tion from the cases of principled value conflicts which matter between
more or less systematic and autonomous doctrinal systems’, instantiated
‘in its purest form by religious or other doctrinal truth claims’
(March 2009: 34, 31). The turn to non-Western religions and philoso-
phies here is animated by a specific motivation to identify areas of moral
disagreement between relatively ‘autonomous systems of argumentation’,
and most importantly to adjudicate in instances of moral conflict, to seek
‘plausible grounds for consensus in other traditions’ (March 2009: 38).
This kind of normative comparative political thought, according
to March, includes and subsumes what is valuable ‘in the weaker form
of comparative political thought, namely the “diagnostic” element of
examining the contours of disagreement between traditions and the
“appreciative” element of demonstrating the diversity of other traditions’.
278 Chapter 13

As a rejection of the term ‘comparative political thought’ simply as

a proxy for non-Western political thought, and as a call for thinking
seriously about why and what we should compare as political theorists,
March (2009) offers a welcome perspective. As a prescription for how
CPT should be done, however, there are reasons for caution. First, the
approach treats different religious traditions as developed within self-
referential cultural systems. This was a feature of religious studies of the
colonial era that has been debunked. March can respond by saying that
his is a stylized account of religion and liberal democracy. Normative-
analytic theorists, as we know, often do rely on models that are not meant
to be literal descriptions of society (well-known examples include the
state of nature and the original position). Nevertheless, at the very least,
this needs to be much more explicit in arguments involving religious
traditions, given the role of intellectual, political and economic power in
defining what religion is and who its followers are. Second, perhaps
deriving from its choice of religious doctrines as the units of comparison,
March’s definition of CPT stipulates unnecessarily demanding criteria
for comparison, notably that comparison should be between units that are
more or less autonomous or mutually exclusive. This relies on an
idealized model of comparison of distinct objects, which has long been
deconstructed in the practice of comparative social and humanities-
oriented disciplines (literature, art, philosophy, religion and politics)
that reject the separateness of the objects of comparison as artificial.
Third, it is simply not the case that comparison in the normative-
analytic mode can subsume without remainder what is valuable in histor-
ical, critical and interpretive forms of political thought (Bajpai 2011).
Normative-analytic theorists have typically wielded highly reified notions
of tradition and culture, whereas the central thrust of historical, inter-
pretive and critical approaches has been to demonstrate the internal
complexity and diversity of unified wholes as well as the dialogical devel-
opment of traditions, cultures and ideas. As such, while the construction
of normative-analytic theory is undoubtedly one important purpose that
a comparative political theorist might pursue, to elevate its requirements
as the standard for all endeavour in the field is to limit the scope of CPT.

3.2 Historical CPT

Historical CPT is by far the most prevalent approach. It comprises
impressive scholarship on key non-Western thinkers, schools of thought
and traditions. We distinguish historical approaches not by the time
periods of the studies (e.g. prior to 1945), but by their methodologies.
Historians of non-Western political thought, like their counterparts
Comparative Political Thought 279

working on European thought, have offered different answers to the

question of how we ought to interpret historical texts from the non-
Western world, which kind of texts we should focus on, how we should
define the appropriate context that is relevant for understanding texts and
how the interpretations of past thinkers can instruct us about current
intellectual and political problems.
What these diverse approaches share is that analysis is driven primarily
by an interest in the recuperation or retrieval of lost or misunderstood
meanings of concepts (e.g. jihad), thinkers (e.g. Islamists such as Qutb
and Mawdudi) and traditions (e.g. Confucianism). Historical interpreta-
tion is of course informed by normative aims, and in some cases expla-
natory intentions, and we can see the family resemblance between certain
normative and historical projects (Ackerly 2005). However, the central
thrust of analysis undertaken is to foster greater appreciation of the
intricacy and sophistication of non-Western thinkers, texts and traditions,
and thereby to enrich and possibly transform the repertoire of political
theory.
Historical approaches in CPT have advanced research methods in
political theory in important respects. First, with regard to the sources
of political thought, political theorists have conventionally focused on
prestigious texts and great thinkers, whereas historians of ideas and
theorists of ideology have highlighted the significance of more mundane
sites of political thought – political pamphlets, propaganda pieces, politi-
cians’ speeches. Students of non-Western political thought have also
suggested that an adequate appreciation of these traditions requires us
to examine not just verbal knowledge articulated in texts and speech, but
also knowledge that is implicit in non-verbal expressions of ritual,
painting, music and dance (Jenco 2007), as historians of pre-modern
periods have also emphasized. Indeed, games, graphics and new media
technologies may ‘hold expressive potentials for political ideas that a more
“conventional” treatise cannot express’ (Goto-Jones 2011: 107).
A second contribution of historical approaches relates to what the
appropriate contexts are for comparative inquiry into non-Western
thought. Theorists have offered two standpoints, one that emphasizes
resemblance, and the other, difference. In the first case, scholars have
located concepts and thinkers of non-Western thought in interpretive
contexts familiar to students of Western political thought. This can
involve, for instance, the recuperation of the histories of concepts such
as democracy and civil society in Arab thought (Browers 2006), of
secularism and liberal ideologies in India (Bajpai 2002, 2012). Roxanne
Euben’s work is a leading example of this strategy. In Enemy in the Mirror,
she argues that ‘Qutb is not a critic of modernity per se – for he views
280 Chapter 13

technologies and scientiﬁc achievements as desirable – but an opponent

of post-Enlightenment rationalism.’ She argues that the ethico-political
worldview of Qutb and other Islamic fundamentalists is best understood
when placed alongside Western critics of Enlightenment rationality such
as Hannah Arendt, Charles Taylor and Alasdair MacIntyre (Euben 1999:
155). Reframing Islamist thought as substantively similar in its formula-
tions to internal Western critics of modernity undermines popular
stereotypes of Islamism as a foreign, extremist and archaic ideology, at
the same time as serving to highlight currently marginal strands within
Western political thought. More generally, the excavation of parallels
between non-Western and Western thought undermines inﬂuential
claims of a clash of civilizations between Islam and the West, or the
‘West and the Rest’. A key challenge its practitioners face, however, is
to mitigate the pitfalls of using Western conceptual frameworks for the
reconstruction of non-Western thought, and thereby imposing Western
categories and frames on non-Western traditions and practices (for
a discussion of the pitfalls and how these can be mitigated, see Godrej
2011, chapter 2).
The second strategy addresses this risk by emphasizing difference. Some
historians of non-Western thought have argued that global political
theory requires us to work primarily from within the terms and
approaches internal to non-European traditions, attempting to jettison,
at least temporarily, the familiar categories of Western political theory.
Leigh Jenco cogently articulates a case for deriving methods for the study
of texts from within the scholarly traditions in which these are embedded
(Jenco 2007: 745). Jenco points out that the ‘frames of inquiry’ of cross-
cultural theorists have remained ‘beholden to modern Western epistemo-
logical debates’ (2007: 745). Proper understanding of non-Western
thought requires not just linguistic knowledge and hermeneutic
sensitivity, but also locating it within the distinct modes of knowledge
production and transmission, and perhaps a willingness to accept ‘foun-
dational hierarchical premises of nondemocratic worldviews’ as they are
(Jenco 2007: 744). The emphasis on cultural distance also informs the
cosmopolitan method advocated by Farah Godrej, in which the theorist
ﬁrst adopts the language and cultural experience associated with the
intellectual and lived tradition of the text, then tries to articulate its
meanings to an audience that has not undergone these existential and
experiential transformations (Godrej 2011). Godrej’s cosmopolitan
method develops Dallmayr’s dialogic approach, which similarly calls for
a comparative method that entails experiential transformation and an
ability to write from the lived experience of more than one tradition
(Dallmayr 1999, 2004).
Comparative Political Thought 281

The emphasis on cultural difference can serve to highlight the ways in

which non-Western thought can potentially extend or displace Western
political theory in ethical and epistemological terms. In many cases,
considerable linguistic, theoretic and ethnographic skills are required
for immersion in a different tradition and the translation of its resources
for Western audiences. The main pitfalls for advocates of the distancing
approach involve how to avoid reinstating an East-West chasm under
a different guise. The attempt to step outside Western concepts and
categories poses two challenges. One is empirical: given historical cross-
cultural engagement, a strong separation between so-called Western and
non-Western traditions is implausible (for a response, see Godrej 2011:
14). The second is the theoretical corollary to the first: the commitment to
working within a tradition reifies the boundary of the tradition a priori.
In fact, the tension between resemblance and difference is a productive
tension for CPT. Instead of treating this as a problem that needs to be
resolved or transcended, a fruitful stance is to treat it as a necessary quality
of the field. The tension is productive methodologically because it
reminds us to self-reflect on the concepts, categories and norms that
define our epistemological perspective. The tension also offers multiple
intellectual contexts for the reconstruction of political thought.
In both normative-analytic and historical CPT discussed thus far, the
practice of comparative political thought has focused overwhelmingly on
key thinkers, exemplary texts and elite intellectual traditions (see also
Thomas 2010). This has served to challenge the Eurocentrism of political
theory in important respects, demonstrating, for instance, that Asia,
Africa, the Middle East and Latin America have produced thought that
matches Western thought in terms of its sophistication and creative
power, and has the capacity to extend its moral and political scope.
A key substantive contribution of these approaches has been to bring
metaphysics back to the study of political thought (Goto-Jones 2011).
CPT scholarship demonstrates that political and metaphysical questions
pursued in separate domains in post-Enlightenment thought are deeply
intertwined.
Nevertheless, the focus on semi-autonomous traditions and key thin-
kers underestimates the field’s potential to expand the frontiers of main-
stream political theory. In the next two sections we outline two
approaches that go beyond the traditional preoccupation of political
theorists with individual thinkers and elite traditions, to engage with
how these are inhabited by people in their lived practices. Whereas
normative-analytic approaches and historical approaches have tended
thus far to emphasize ‘non-Western’ sources, as well as the Western and
non-Western dichotomy as the telling axis of comparison (for an
282 Chapter 13

exception, see Goto-Jones 2011), the other two approaches seek to break
down that dichotomous construction of ‘comparison’ and the focus on
‘traditions’ that has often accompanied this.

3.3 Interpretive CPT

Interpretive theorists challenge the view shared by the other approaches,
that the primary purpose of political theory is normative or prescriptive,
that of advancing visions of the right or the good (Freeden and Vincent
2013). Instead, in interpretive social science writings, the political-
theoretic task of conceptual reconstruction serves also to theorize real-
world political processes, and to advance explanations of political
outcomes (Bevir and Rhodes 2002; Bajpai 2011). Interpretive social
science approaches have often been conﬂated with historical approaches,
but these are distinct. Interpretive theorists use social scientiﬁc methods –
qualitative data and qualitative methods of analysis (see Yanow and
Schwartz-Shea 2013) – to provide detailed empirical accounts of mean-
ings and ‘thick description’, to use Geertz’s well-known term (Geertz
1973; Taylor 1985). Data sources can include constitutional and policy
debates, interviews, participant observation, focus groups and ethnogra-
phy. Methods of analysis can include discourse analyses of elite and
colloquial texts (Bajpai 2011; Tripp 2013), as well as political ethnogra-
phies (Mahmood 2005; Wedeen 2008).
Like historical CPT scholars, interpretive social scientists too seek to
reconstruct the meanings of ideas in particular contexts – what democ-
racy, social justice, civil society mean to people in particular times and
places. However, there are important differences with historical
approaches. First, in interpretive CPT, the focus is not on the thought
of exemplary thinkers or ‘innovating ideologists’ (Bajpai 2011), on the
appreciation of the theoretical and political creativity of individuals such
as Qutb or Gandhi. Instead, interpretive social scientists focus on what
Lisa Wedeen terms meaning-making practices – discursive, rhetorical
and performative. Practices pertain to what agents do, and how this
interacts with language and other symbolic systems (Wedeen 2002:
714). Thought practices and meaning construction are not pursued
here in the form of the ideas, beliefs or writings of individuals, as is
common in historical approaches, but, rather, as embedded and framed
within collective actions, both everyday and extraordinary (e.g. of com-
pliance or protest) as characteristic of anthropological writings
(Mahmood 2005; Wedeen 2008). For instance, Charles Tripp’s account
of ideas of political participation in the Middle East is based more on their
articulation in repertoires of collective action and protest than on the
Comparative Political Thought 283

writings of individuals (Tripp 2013). Tripp distinguishes conceptions of

participation that derive from different types of political struggles – for
a nation-state against Western rule, for ‘dimuqratiyyat al-khubs [democ-
racy of bread]’ through direct action and occupation of public spaces, and
for the preservation of an Islamic community through ‘observance of the
shari’a’ (Tripp 2013: 91, 95–7). Humeira Iqtidar’s account of shifts in
Islamist political imaginaries in Pakistan goes beyond the focus on found-
ing ideologues such as Maududi, to examine the narratives of ordinary
activists and supporters of the Jama’at-e-Islami engaged in social and
humanitarian work (Iqtidar 2011).
Second, interpretive CPT can be explanatory in ways that are distinct
from historical CPT. Historians of political thought have also engaged
explanation, but usually in terms of explicating why ideas assumed the
form that they did at a given historical moment. Occasionally, links are
posited between ideas and political outcomes. For instance, Roxanne
Euben’s study of Sayyid Qutb pursues a ‘deeper, richer, fuller account
of Islamist ideas’, also because demonstrating ‘the intellectual coherence
and depth of fundamentalist ideas makes explanations of Islamic
fundamentalism more causally adequate’ (1999: 156). However, in
most historical approaches, the further task of identifying and establishing
such links between thought and action in speciﬁc contexts remains unad-
dressed. By contrast, in many instances of interpretive CPT, the recon-
struction of thought focusses directly on improving explanations of
particular political phenomena. How are religio-political identities (e.g.
being Muslim, Hindu) produced, and why do these gain salience in
certain contexts, and not others? How is political compliance produced,
and why do ‘some political ideologies, policies, and self-policing strate-
gies work better than others’? How do actors’ perceptions of what
‘democracy’ and ‘religion’ mean affect political outcomes (Wedeen
2002: 714)? For instance, Rochana Bajpai’s account of conceptions
of secularism, social justice, democracy, national unity and develop-
ment in India based on the practices of political argument in con-
stitutional and legislative debates seeks to delineate the distinctive
features of these concepts in India, and also to improve explanations
of policies of minority rights and afﬁrmative action at critical junc-
tures (Bajpai 2011).
Ideational explanation here has two elements. First, our explanations
of political outcomes are necessarily incomplete without a grasp of the
normative resources available to actors, just as they would be without
knowledge of other resources – economic, institutional – available to
them (Bajpai 2011). Here, ideas or norms can serve as an explanatory
variable. For example, Matthew Nelson has shown how the shift from
284 Chapter 13

custom to shari’ah radically altered the pattern of local politics in Pakistan

(Nelson 2011). Second, ideational practices (e.g. debates in institutional
as well as non-institutional fora) serve as a lens for viewing political
processes, for gaining better descriptions and thereby explanations of
political change, in particular of shifts in relations of power (Asad 1993;
Wedeen 2002: 714). For instance, the centralization of power in a polity
at a given historical moment and its decentralization at another are
manifested not necessarily, or only in the nature of party competition,
but also in patterns of political argument, which a conceptual analysis of
policy discourse reveals (Bajpai 2011).
Interpretive approaches have the potential to advance CPT in impor-
tant respects. First, the focus on practices as the units of analysis opens up
the category of ‘tradition’, which has often privileged elite sources and
reinforced cultural boundaries. It offers the prospect of more nuanced
assessments of similarity and difference across different traditions, and
over time within a tradition. Second, interpretive CPT approaches typi-
cally focus on the present, bringing attention to bear on the recent
political experience of Asia, Africa, the Middle East and Latin America
that has remained relatively neglected in historical and normative CPT
thus far. Third, interpretive CPT extends the role of political theory in
political science, specifically in relation to ideational explanations of
political outcomes, which remain underdeveloped in social science scho-
larship (see Bajpai and Brown 2013).
Nevertheless, one potential problem with interpretive approaches
is that in expanding the role of CPT to include political explanation that
is conventionally associated with political science, it risks losing what is
distinctive in political theory. Its advocates reply that using the tools of
political theory to conceptualize and theorize political phenomena and
processes enhances the real-world relevance of political theory. This latter
interpretation is in keeping with the historical emergence of the discipline
in which political theory served as the reflective dimension of political
science, helping the field define its questions and concepts (Wolin 1969;
Gunnell 2010). Furthermore, within political theory, engagement with
empirical politics on the part of political theorists is a growing trend.
Here, the task of political theory is not seen just in terms of providing
solutions to our current moral predicaments, but also as clarifying what
these are, in helping give ‘form to emergent realities that otherwise remain
beyond our ken’ (cf. Wolin 1969; Isaac 1995; Kaufman-Osborn 2010:
668), and thereby relocating political theory within the realm of political
science (Freeden 2007; Gunnell 2010).
Comparative Political Thought 285

3.4 Critical CPT

In the Marxist tradition of political theory, critical CPT connects to the
‘struggles and wishes of the age’ (Marx 1967). Critical CPT identiﬁes the
dissenting strands within elite traditions of political theory and the non-
elite voices within the lived experience of political thought.
Methodologically, critical CPT is the study of elite and non-elite actors,
of canonical and non-mainstream texts, of texts and the actions and aspira-
tions of those in struggle. Some of these actors can be marginalized in
global politics or by local politics in struggles with elites. Globally and
locally, marginalization differs whether due to race, religion, ethno-
nationality, language, sexuality or perceived sexuality.
Christine Keating situates her argument about the meaning of the
social contract in the context of the decolonization of India (Keating
2011). Luis Cabrera locates his argument about global citizenship in
the actions of global migrants, citizens policing the borders and those
engaged in humanitarian aid in the borderlands (Cabrera 2010). Brooke
Ackerly draws on the insights of women’s rights activists from around the
world to articulate a theory of human rights from the practice of struggle
for rights (Ackerly 2008). Kim places his argument about Confucian
democracy and law in the context of contemporary South Korea politics
over the meaning of democracy and freedom (Kim 2014).
Sometimes called grounded political theory, these approaches are quin-
tessentially question-driven inquiry, framing their questions around the
political struggles of non-elites, that is, those who are not privileged in
their contemporary political contexts. This approach seeks to broaden the
history of ideas at stake in the political struggle to include not just the
ideas of the winners, but also of those who struggled for a different
political practice. The actions and aspirations of those in struggle are
treated as text on a par with the texts inherited by elites. It also treats
activists as political theorists (Cabrera 2010; Keating 2011: 114), and
applies the insights gained by studying the struggles in one context (India)
to the struggles in another (Keating 2011), to make more general claims
about speciﬁc concepts and their use (Ackerly 2008; Beltrán 2009;
Cabrera 2010), and to challenge familiar conceptualizations.
While critical, this view entails a general claim: that it is the study of
ideas and concepts that are revealed in the struggles of politics, not just in the
ideas and concepts that emerge victorious from those struggles. Certain
substantive universals may emerge from such inquiry. But these claims
proceed with a sceptical scrutiny for the potential to false essentialisms,
false universalisms and exclusions. These can be made through the intent
to include or broaden the perspectives being considered. At their best,
286 Chapter 13

critical grounded approaches to CPT work in alliance with non-elites in

political struggle; this is a methodological alliance for understanding the
struggle, not an ideological alliance of researcher and political actors.
One interesting potential problem with critical approaches is that they
can be mired in discursive politics. If these politics get disassociated from
the underlying problem or used to silence or deny the voice of some
participants in the struggle (Mackey 2005; Rothschild, Long and Fried
2005; Fricker 2007; Medina 2012), then critical CPT does not lead to
political or theoretical insights more fruitful than the more elite-text-
based historical approaches to CPT. The critical grounded approach
requires either an empirical component or triangulation across compara-
tive methods.
Finally, critical CPT approaches use normative, historical and
interpretive methods. What distinguishes critical CPT is its authors’
normative commitment to those contemporary political struggles. In the
following section, we turn to the discussion of methods – speciﬁc tools –
that multiple methodologies may use to effect the author’s goals.

4 How to Do CPT
In a CPT project, as in other subfields of politics, methodologies and
methods need to be chosen with respect to a particular question. CPT
scholars often use a mixed methodology, perhaps beginning in one meth-
odology and revising using another. In this light, it is best to think of the
four CPT methodologies as providing an initial guiding framework.
As with empirical social science, triangulation across methodologies and
methods can be more fruitful than relying on one. The key to avoiding the
potential for epistemological imperialism in political theory is to be ques-
tion-driven, attentive to methodological pluralism even if your own methods
tend towards one strand of the field, attentive to the potential of any
method to be self-centric due to the building blocks of scholarship, and
self-reflective about the best execution of the methods of your selected
approach or approaches. These are ‘meta-methods’ if you will. Many,
maybe even all of these, are or should be part of any political theory
practice. Post-colonial, feminist and other critical approaches have
made similar arguments. Nevertheless, we offer a CPT perspective on
research practices and outline their concrete instantiation in CPT
scholarship.
(a) Start with your research question rather than methodology
or method. Several scholars have called for more question-driven
research, in contrast with the emphasis on methods and models in poli-
tical theory and political science (on the general point, see Isaac 1995;
Comparative Political Thought 287

Shapiro 2005). CPT joins calls in the discipline for putting the question
before methods. It has the potential to enlarge research questions beyond
those traditionally addressed by political theorists to include those
informed by deep engagement with non-Western sources on the one
hand, and empirical social science on the other.
As with other forms of political theory, a CPT question can be based on
an empirical observation. These can be empirical problems (such as
violence, immigration, climate change, gender oppression and poverty)
that pose challenges for how we live together. Such an empirical problem
can also lead to a theoretical problem that needs to be addressed before
we can think about how we might address the empirical problem (and
others like it). As we saw earlier, CPT scholars have addressed questions
of governance, human rights, democracy and citizenship. A CPT ques-
tion can also come from a new way of reading a text that comes from the
reading of texts in and/or across traditions. In sum, a CPT question can
be grounded in a problem (e.g. poverty), a concept (e.g. human rights) or
texts (e.g. Confucian and Neo-Confucian texts). Be aware of the fact that
your question will inﬂuence your methodological commitments.
(b) Know your ontological perspective and be open to its revi-
sion. By your selection of a research project, you reveal to yourself and
the world the ontological perspective behind the inquiry. Some
differences among theorists can be ontological, but manifest themselves
methodologically. For example, Andrew March argues that the important
differences across historical traditions lie in the differences in their world-
views. In fact, March has a worldview (an ontological perspective) about
the differences in worldviews. By contrast, Roxanne Euben argues that by
familiarizing ourselves with seeming differences in worldviews, we can
gain better self-understanding and an appreciation that differences are
not as different as we might have thought. She has an ontological per-
spective on the inter-comparability and inter-compatibility of world-
views. In Enemy in the Mirror, she argues that ‘Qutb’s work must be
understood as a “dialectical response” to rationalism and
Westernization’, and that he is ‘participating in a conversation that we,
as Western students of politics, not only recognize, but in which we
participate’ (Euben 1999: 155).
Consider the ways in which the tensions among competing readings or
interpretations and seeming ‘discontinuities’ or irreconcilable differences
across time and geography reveal not just intra-disciplinary methodolo-
gical disagreements, but ontological differences. If you are doing CPT
well, you challenge yourself to revisit your ontological perspectives
throughout the research process.
288 Chapter 13

(c) Know your methodology and be open to its development.

Question-driven research is in principle methodologically plural at the
moment of inquiry. None of the methodologies discussed earlier requires
that you use just one method. Each exempliﬁes a particular combination of
methods. For instance, Euben in the example cited previously also uses
interpretive and critical methods to some extent when she situates her
reading of Qutb in the context of contemporary politics of his reception by
Sunni and Western audiences.
Likewise, the same method can be used across methodologies. For
example, historical and interpretive social scientists use textual analysis;
however, interpretivists usually try to supplement this ‘with more ethno-
graphic modes of analysis (interviews, observation, etc.) so as to give more
widespread accounts of the beliefs found among the different actors
involved in a practice’.2
In choosing your methodology, ask yourself what your main purpose in
undertaking CPT is. Is it the retrieval or recuperation of particular non-
Western thinkers or debates? If so, historical methodologies are likely to
be most appropriate. Is it to respond to contemporary challenges that
affect people across the globe (e.g. climate change) from the standpoint of
the experience of those most vulnerable in Asia and Africa? If so, critical
or normative methodologies are likely to be more suitable. Alternatively,
your main interest may be the conceptualization and theorization of new
forms of political action among elites and/or subalterns in Asia, Africa or
the Middle East, for which interpretive CPT methodologies are likely to
be most useful. Each methodology embodies a particular priority among
the methods it deploys in order to resolve conﬂicts and ascertain the
validity of conclusions.
(d) Know which methods you will need and be open to multiple
methods. Consider the possible methods and select the right combina-
tion for your question. As with other areas of research, in CPT the use of
multiple methods across different methodologies has yielded rich
insights.
For example, conceptual analysis is typically used in analytic meth-
odologies to construct normative arguments (see Olsthoorn’s chapter in
this volume and the discussion of the normative-analytic approach
earlier), including arguments about the possible kinds of desirable poli-
tical change (Angle 2005; Kim 2014). However, it can also be used in
interpretive CPT, where conceptual reconstruction serves to advance
explanations of political outcomes and political change (e.g. Bajpai
2011). Interpretive CPT offers an interstitial space between political

2
Mark Bevir, personal communication, May 2015.
Comparative Political Thought 289

theory and political science, where the tools of political theorists are used
for improving social science explanations.
Critical comparative political theorists have analysed non-elite texts and
even treated as texts the arguments that are embedded in the political actions
of non-elites who are seeking alternative political arrangements. Critical
CPT methods include observation of contemporary struggles (Beltrán
2009; Cabrera 2010; Clifford 2012; Forman and Cruz forthcoming), critical
rereading of historical struggles (Keating 2011) and even interviews with
actors in the struggle (Ackerly 2008). Sungmoon Kim’s Confucian
Democracy in East Asia utilizes historical, interpretive and critical methods
and normative-analytic argument in setting out the import of his project
(Kim 2014).
Be aware that the findings from different methods may conflict.
Methodological location and triangulation (see later) can help with the
resolution of such conflicts.
(e) Be aware of your own boundaries. It is important to recognize
that as a comparative political theorist, you both construct and decon-
struct boundaries as you work. Attentiveness to that construction and
reconstruction is a feature of CPT best practice.
Consider, for example, the political and theoretical debate in the 1990s
about ‘Asian Values’ and their (in)commensurability with political equality,
individual freedom and rights. In the Asian Values debate, some partici-
pants emphasized the distinctiveness of East Asian cultures (Zakaria 1994;
Bell 2000). Drawing on textual analysis and cross-cultural comparison,
they developed a reified dichotomy between Western and Asian traditions.
Participants in that same debate challenged those readings of historical
texts and offered competing interpretations of ‘the’ cultural tradition
(Kim Dae 1994; Chan 1997). The debate revealed great differences within
each of these intellectual and cultural groupings. The diversity within those
broad civilizational categories – and their political import for charting
possible ways of thinking about what ethical responsibility requires in the
face of contemporary challenges – is obfuscated by the overgeneralized
characterization of and sharp delineation between the dialogue partners.
(f) Acquire requisite language skills. CPT methods typically, but
not necessarily, require competency in the language of the texts.
Depending on the research question, the availability and quality of trans-
lations and knowledge about the politics and epistemology of a text’s
translators, some CPT can be done without fluency.
(g) Read texts closely and contextualize these. All CPT makes use
of close textual reading. Whether your reading is typical or a nuanced
rereading, be attentive to alternative readings of the texts (see also Blau’s
chapter in this volume). Debates from the time period, contemporary
290 Chapter 13

secondary literature and political history of the context can enhance the
reading. Whereas the focus in reading texts is often the substance of their
arguments, CPT scholars can also read texts for their insights into how to
do political theory (e.g. Jenco 2007).
(h) Consider the broad range of available sources. When your
research question relates to broad categories – such as culture or religion –
contextualize your account of any particular text or other evidence with
reference to a broad range of sources. Remember (or reconsider) your
ontological commitments when you identify your sources.
From a critical CPT perspective, it is especially important for you to
consider the status of the authors of your texts at the time and in history.
Much CPT broadens our sources, but you may not be able to include
non-elite sources, those either lost to time because they did not generate
texts that are available today or omitted because their texts are not
considered ‘theory.’
For example, in East Meets West, Daniel Bell broadens what counts as
a text for political theory by taking the political ideas of two leaders in the
Asian Values debate of the 1990s and articulating their views in the form
of ﬁctionalized characters. However, the ﬁctionalized protagonists of the
dialogue that he constructs between East and West are as elite as the
political leaders on whose ideas the text is based (Bell 2000). Neither
protagonist is engaged in a struggle for recognition and enjoyment of his
own human rights. This is a critical CPT assessment of Bell’s sources.
(i) Pick your qualitative methods carefully. Political practice can
be a source of CPT ‘text’ for analysis. Michael Freeden, in his call for
explicit methods for CPT, urges theorists to focus more on ‘real world
forms of political thinking’ rather than great thinkers and texts that are
unrepresentative of their contexts (2007: 2; see also Bajpai 2011). Godrej
(2011) calls on CPT work to be anthropological or ethnographic.
However, such work can also be qualitative without being ethnographic.
For example, Ackerly (2001) looks at the human rights practices of
activists to complement what they say about their work. Cabrera (2010)
looks at the actions of migrants and those near and crossing borders.
Be aware too that textual and written sources also encode practices and
ritual forms – the latter are not just to be found in the realm of ‘behaviour’.
(j) Attend to sources of bias. A narrow source list can be one source
of bias, but there are others. Because each methodological approach has
strengths and weaknesses, be attentive to and make explicit the biases in
your framing of the comparative project. For instance, does it frame
Islamic, Indian or Chinese traditions in terms of their orthodox variants,
or even their dominant heterodox strands? What alternative sources pro-
vide a counter-reading of the thinkers and traditions that are its focus?
Comparative Political Thought 291

Have you constructed a dichotomous juxtaposition of two readings of

a text? Are there others whose nuance is lost by your heuristic? In what
ways are your empirical sources limited or limiting?
(k) Triangulate. Even if you use only one methodological approach,
use your awareness of the other methodologies to create a reflective
dialogue with yourself about the possible other approaches to studying
your question. Consider other approaches. What would the approach
consider or recommend, why is it tempting, and why is it less suited
your question? Following Goodin’s advice in this volume, show your
work. The standard of showing your work is more detailed in a book-
length treatment. However, the excuse of space for not treating a serious
criticism or alternative reading is not intellectually satisfying. Space con-
straints may justify not pursuing a mixed methodology approach, but
again, it is a weak response to not reflecting on the insights such an
approach might offer your project.
(l) Consider co-authorship. Given the specialization of debates
within political theory, the voluminous scholarship on the Middle East,
Asia, Africa and Latin America, not to speak of the desired linguistic
skills, serious comparative work requires considerable time and effort.
Furthermore, as we have discussed and demonstrated in our co-
authorship of this chapter, behind all CPT methodologies is an
epistemological commitment to broadening the scope of insights that
are brought to bear on a question. No matter your skills, the scale of
comparative work and the scope of insight can be broadened through co-
authorship.
(m) Be humble. When writing from a position of privilege, and most
political theorists do, humility vis-à-vis what you can know is appropriate.
Many of the ‘how-to’ points make reference to limitations, such as limited
access to historical knowledge, disproportionate access to elite texts and
limited access to the lived politics of those in struggle. In addition, use
humility to reflect on the privileges of being able to do this work and the
intellectual obligations of this privilege.

5 Conclusions
In the twenty-ﬁrst century, the ﬁeld of political theory and comparative
political thought is global, not in the sense that any articulation of the
complex web of concepts would be globally agreed to or that such agree-
ment should be the goal, but rather that our interlocutors in this endea-
vour are not predetermined by our training, experience, geography or
imagination, but may come from any place, time or family of inquiry.
292 Chapter 13

CPT can help us do political theory better. The field’s internal discus-
sions of methodologies and method can help its practitioners – novice and
experienced – do their work better. One day the modifier ‘comparative’
will become associated with this time in the historical development of our
discipline because in fact what we learn from how we do comparative
inquiry improves how we do political theory (Dallmayr 2004; Euben
2006: see especially chapter 2; March 2009: 536–7; see also Leader
Maynard’s chapter in this volume). CPT enables us to deepen our reflec-
tions about what we mean by ‘we’ when ‘we’ as theorists engage in an
ethical and political reflection.
Political, economic and social empirical problems with normative
import are the important questions for political theorists. Because
many of these recur, the history of political thought is a likely source
for a wealth of reflective insights. Because these issues have been
relevant in the world and over time, the historical intellectual tradi-
tions that may provide insight may come from anywhere. Because
these issues are pressing now, contemporary theorists around the
world should draw on each other’s reflective insights in order to
broaden our understanding of the web of relevant concepts and help
clarify our articulations of them. The vastness of these ambitions
extends beyond the capacity of any individual’s life work. Therefore,
political theory relevant to the significant challenges posed by these
ambitious normative puzzles needs to be a collaborative enterprise.
Taken together, the range of methodological forms that CPT takes
enables political theory to ever improve its contributions to the study
and practice of politics.
CPT methods build a shared, though not necessarily common platform
across multiple domains of knowledge production. They challenge false
universalisms and dominant Western stereotypes. They situate political
theory in multiple contexts of resemblance and difference and expand the
notion of text to include debates and action. CPT is one of the subfields in
political theory that engages in constructive criticism of the ways in which
political theory can perform global politics, not just write about it. CPT
methods push political theory to become increasingly self-aware about
that possibility.

Acknowledgements
We are grateful to participants in the London Comparative Political
Thought seminar, and to Adrian Blau, Humeira Iqtidar, Leigh Jenco,
Sungmoon Kim and Tejas Parasher for their comments on earlier
versions.
Comparative Political Thought 293

References
Ackerly, Brooke, 2001. ‘Women’s human rights activists as cross-cultural theor-
ists’, International Feminist Journal of Politics 3, 311–46.
Ackerly, Brooke, 2005. ‘Is liberalism the only way toward democracy?
Confucianism and democracy’, Political Theory 33, 547–76.
Ackerly, Brooke, 2008. Universal Human Rights in a World of Difference. Cambridge
University Press.
Ackerly, Brooke and Jacqui True, 2010. Doing Feminist Research in Political and
Social Science. Palgrave Macmillan.
Ackerly, Brooke and Jacqui True, 2012. ‘Methods and methodologies’, in
Georgina Waylen, Karen Celis, Johanna Kantola and Laurel Weldon, eds.,
The Oxford Handbook of Gender and Politics. Oxford University Press, 109–33.
Angle, Stephen, 2002. Human Rights and Chinese Thought: A Cross-cultural Inquiry.
Cambridge University Press.
Angle, Stephen, 2005. ‘Decent democratic centralism’, Political Theory 33,
518–46.
Angle, Stephen, 2012. Contemporary Confucian Political Philosophy. Cambridge:
Polity.
Asad Talal, 1993. Genealogies of Religion: Discipline and Reasons of Power in
Christianity and Islam. Baltimore, MD: Johns Hopkins University Press.
Bajpai, Rochana, 2002. ‘The conceptual vocabularies of secularism and minority
rights in India’, Journal of Political Ideologies 7, 179–97.
Bajpai, Rochana, 2011. Debating Difference: Group Rights and Liberal Democracy in
India. Oxford University Press.
Bajpai, Rochana, 2012. ‘Liberalisms in India: a sketch’, in Ben Jackson and
Marc Stears, eds., Liberalism as Ideology: Essays in Honour of Michael Freeden.
Oxford University Press, 53–76.
Bajpai, Rochana and Graham Brown, 2013. ‘From ideas to hegemony: ideational
change and afﬁrmative action policy in Malaysia, 1955–2010’, Journal of Political
Ideologies 18, 257–80.
Bell, Daniel, 2000. East Meets West: Human Rights and Democracy in East Asia.
Princeton, NJ: Princeton University Press.
Bell, Daniel, 2006. Beyond Liberal Democracy: Political Thinking for an East Asian
Context. Princeton, NJ: Princeton University Press.
Bell, Daniel, 2008. China’s New Confucianism: Politics and Everyday Life in
a Changing Society. Princeton, NJ: Princeton University Press.
Beltrán, Cristina, 2009. ‘Going public’, Political Theory 37, 595–622.
Bevir, Mark and R. A. W. Rhodes, 2002. ‘Interpretive theory’, in David Marsh and
Gerry Stoker, eds., Theory and Methods in Political Science. Basingstoke: Palgrave,
131–52.
Browers, Michaelle, 2006. Democracy and Civil Society in Arab Political Thought:
Transcultural Possibilities. Syracuse University Press.
Cabrera, Luis, 2010. The Practice of Global Citizenship. Cambridge University
Press.
Chan, Joseph Cho Wai, 1997. ‘An alternative view’, Journal of Democracy 8, 35–48.
294 Chapter 13

Chan, Joseph Cho Wai, 2009. ‘Is there a Confucian perspective on social justice?’,
in Takashi Shogimen and Cary Nederman, eds., Western Political Thought in
Dialogue with Asia. Lanham, MD: Lexington Books, 261–77.
Chan, Joseph Cho Wai, 2015. Confucian Perfectionism: A Political Philosophy for
Modern Times. Princeton, NJ: Princeton University Press.
Clifford, Stacy, 2012. ‘Making disability public in deliberative democracy’,
Contemporary Political Theory 11, 211–28.
Dallmayr, Fred, ed., 1999. Border Crossings: Toward a Comparative Political Theory.
Lanham, MD: Lexington Books.
Dallmayr, Fred, 2004. ‘Beyond monologue: for a comparative political theory’,
Perspectives on Politics 2, 249–57.
Euben, Roxanne Leslie, 1999. Enemy in the Mirror: Islamic Fundamentalism and the
Limits of Modern Rationalism. Princeton, NJ: Princeton University Press.
Euben, Roxanne Leslie, 2006. Journeys to the Other Shore: Muslim and Western
Travelers in Search of Knowledge. Princeton, NJ: Princeton University Press.
Forman, Fonna and Teddy Cruz, forthcoming. ‘Global justice at the municipal
scale: the case of Medellín, Colombia’, in Thomas Pogge and Luis Cabrera, eds.,
Institutional Cosmopolitanism. Oxford University Press.
Forst, Rainer, 2001. ‘Towards a critical theory of transnational justice’,
Metaphilosophy 32, 160–79.
Freeden, Michael, 1996. Ideologies and Political Theory: A Conceptual Approach.
Oxford: Clarendon Press.
Freeden, Michael, 2007. ‘The comparative study of political thinking’, Journal of
Political Ideologies 12, 1–9.
Freeden, Michael and Andrew Vincent, eds., 2013. Comparative Political Thought:
Theorizing Practices. New York: Routledge.
Fricker, Miranda, 2007. Epistemic Injustice: Power and the Ethics of Knowing. Oxford
University Press.
Fukuyama, Francis, 1992. The End of History and the Last Man. New York: Free
Press.
Gaus, Gerald, 1990. Value and Justiﬁcation: The Foundations of Liberal Theory.
Cambridge University Press.
Geertz, Clifford, 1973. The Interpretation of Cultures: Selected Essays. New York:
Basic Books.
Godrej, Farah, 2011. Cosmopolitan Political Thought: Method, Practice, Discipline.
Oxford University Press.
Gordon, Jane Anna, 2014. Creolizing Political Theory: Reading Rousseau through
Fanon. Bronx: Fordham University Press.
Goto-Jones, Chris, 2011. ‘A cosmos beyond space and area studies: toward com-
parative political thought as political thought’, boundary 2 38, 87–118.
Gunnell, John, 2010. ‘Professing political theory’, Political Research Quarterly 63,
674–79.
Herzog, Don, 1985. Without Foundations: Justiﬁcation in Political Theory. Ithaca,
NY: Cornell University Press.
Huntington, Samuel, 1996. The Clash of Civilizations and the Remaking of World
Order. New York: Simon & Schuster.
Comparative Political Thought 295

Iqtidar, Humeira, 2011 ‘Secularism beyond the state: the “state” and the “market”
in Islamist imagination’, Modern Asian Studies 45, 535–64.
Isaac, Jeffrey, 1995. ‘The strange silence of political theory’, Political Theory 23,
636–52.
Jenco, Leigh, 2007. ‘“What does heaven ever say?” A methods-centered approach
to cross-cultural engagement’, American Political Science Review 101, 741–55.
Jenco, Leigh, 2011. ‘Recentering political theory: the promise of mobile locality’,
Cultural Critique 79, 27–59.
Kaufman-Osborn, Timothy, 2010. ‘Political theory as profession and as subﬁeld?’,
Political Research Quarterly 63, 655–73.
Keating, Christine, 2011. Decolonizing Democracy: Transforming the Social Contract
in India. Pennsylvania State University Press.
Kim Dae, Jung, 1994. ‘Is culture destiny: the myth of Asia’s anti-democratic
values’, Foreign Affairs 73.
Kim, Sungmoon, 2014. Confucian Democracy in East Asia: Theory and Practice.
Cambridge University Press.
Laborde, Cécile, 2008. Critical Republicanism: The Hijab Controversy and Political
Philosophy. Oxford University Press.
Mackey, Eva, 2005. ‘Universal rights in conﬂict: “backlash” and “benevolent
resistance” to indigenous land rights’, Anthropology Today 21, 14–20.
Mahmood, Saba, 2005. Politics of Piety: The Islamic Revival and the Feminist Subject.
Princeton, NJ: Princeton University Press.
March, Andrew, 2009. ‘What is comparative political theory?’, The Review of
Politics 71, 531–65.
Marx, Karl, 1967. Writings of the Young Marx on Philosophy and Society, ed. Loyd
David Easton and Kurt Guddat. Garden City, NY: Doubleday.
Medina, José, 2012. The Epistemology of Resistance: Gender and Racial Oppression,
Epistemic Injustice, and Resistant Imaginations. Oxford University Press.
Mills, Charles, 2015. ‘Decolonizing Western political philosophy’, New Political
Science 37, 1–24.
Nelson, Matthew, 2011. In the Shadow of Shari, ah: Islam, Islamic Law, and
Democracy in Pakistan. Columbia University Press.
Parel, Anthony, 2006. Gandhi’s Philosophy and the Quest for Harmony. Cambridge
University Press.
Rothschild, Cynthia, Scott Long and Susana Fried, eds., 2005. Written Out: How
Sexuality Is Used to Attack Women’s Organizing. New York: International Gay
and Lesbian Human Rights Commission & The Center for Women’s Global
Leadership.
Shapiro, Ian, 2005. The Flight from Reality in the Human Sciences. Princeton
University Press.
Tan, Sor-Hoon, 2003. Confucian Democracy: A Deweyan Reconstruction. Albany:
State University of New York Press.
Taylor, Charles, 1985. Philosophy and the Human Sciences. Cambridge University
Press.
Thomas, Megan, 2010. ‘Orientalism and comparative political theory’, The Review
of Politics 72, 653–77.
296 Chapter 13

Tripp, Charles, 2013. ‘Acting and acting out: conceptions of political participation
in the Middle East’, in Michael Freeden and Andrew Vincent, eds., Comparative
Political Thought: Theorizing Practices. New York: Routledge, 88–109.
Von Vacano, Diego, 2015. ‘The scope of comparative political theory’, Annual
Review of Political Science 18, 465–80.
Wedeen, Lisa, 2002. ‘Conceptualizing culture: possibilities for political science’,
American Political Science Review 96, 713–28.
Wedeen, Lisa, 2008. Peripheral Visions: Publics, Power, and Performance in Yemen.
University of Chicago Press.
Whitehead, Laurence, 2013. ‘Latin American approaches to “the political”’, in
Michael Freeden and Andrew Vincent, eds., Comparative Political Thought:
Theorizing Practices. New York: Routledge, 40–59.
Wolin, Sheldon, 1969. ‘Political theory as a vocation’, American Political Science
Review 63, 1062–82.
Yanow, Dvora and Peregrine Schwartz-Shea, eds., 2013. Interpretation and
Method: Empirical Research Methods and the Interpretive Turn. Armonk, NY:
M. E. Sharpe.
Zakaria, Fareed, 1994. ‘Culture is destiny: a conversation with Lee Kuan Yew’,
Foreign Affairs 73, 109–26.
14 Ideological Analysis

Jonathan Leader Maynard

1 Introduction
The past two decades have seen a proliferation of academic work on
ideology (Freeden et al. 2013: v). While the study of ideology is older
than this (Larrain 1979; McLellan 1995), recent developments have
made it increasingly appropriate to speak of ‘ideology studies’ or ‘ideolo-
gical analysis’ as a distinct interdisciplinary field of research within the
humanities and social sciences. Researchers of ideology have made sig-
nificant methodological, theoretical and empirical advances, in
disciplines as diverse as political theory (Freeden 1996), intellectual
history (Tully 1983; Skinner 2002b), political psychology (Rosenberg
1988; Jost and Major 2001; Haidt et al. 2009; Jost et al. 2009a), discourse
analysis (Howarth et al. 2000; Fairclough 2010), political science (Knight
2006; Carmines and D’Amico 2015), sociology (Boudon 1989) and
social and cultural studies (Eagleton 1991; Hall 1996; Shelby 2003).
Indeed, the biggest problem facing the contemporary study of ideology
is the fragmentation of work across disciplines – one thing this chapter
attempts to address (for existing interdisciplinary work, see Žižek 1994a;
van Dijk 1998; Freeden 2007; Freeden et al. 2013). But while problems
and lacunae remain, ideological analysis is currently at a high point of
sophistication, diversity and output (for guideline maps to such contem-
porary research, see Norval 2000; Leader Maynard 2013; Leader
Maynard and Mildenberger 2016).
The overwhelming bulk of this work has occurred within research
domains that could loosely be described as empirical. But that is not the
focus of this chapter. I contend that ideological analysis is also important
for the sort of conceptual and normative political theory covered in most
chapters of this volume. Some theorists, notably the leading ideology
scholar Michael Freeden, have argued for the relevance of ideological
analysis to political theory by advocating a broad conceptualization of the
latter: where political theory goes beyond conceptual and normative
analysis to include theoretical reflection about empirical features of

297
298 Chapter 14

politics (Freeden 2005b, 2006, 2008, 2013b: especially chs. 1 and 2).
There are merits to this argument, but this is also not my approach here.
Instead, I’m happy to limit the scope of political theory to a normative
form – though I assume that this encompasses a broad range of ‘ideal’ and
‘non-ideal’ theorizing (Valentini 2012; see too the chapters by Jubb and
by Schmidtz in this volume). I will show how ideological analysis can
support such normative political theory (although my discussion should
be useful for those engaged in empirical work too). In Section 1, I discuss
how to conceptualize ideology, before going on, in Section 2, to explain
how ideological analysis can be valuable for normative political theory,
arguing that political theorists need to attend to it more frequently.
Finally, in the more substantive Section 3, I offer an integrative account
of how to engage in rigorous ideological analysis.
At many points my comments are necessarily cursory. My objective here
is to offer the most practically useful guide to engaging in ideological
analysis for scholars and students. Offering the fullest justiﬁcation for
every piece of advice I give would make it impossible to explain that advice
clearly. Instead, my bibliographical references should serve to link readers
to key texts that examine the corresponding issues in more depth.

2 The Meaning of Ideology

To analyse ideology is to be concerned with two important features of
human beings. First, ideological analysis requires a recognition that
different individuals, groups, institutions or societies are charac-
terized by distinctive idiosyncratic worldviews that meaningfully
shape their political thought and political behaviour.
To understand, explain or predict what they say, think and do, we there-
fore need to identify and study those worldviews. This stands in contrast
to modes of political analysis that present human beings as fundamentally
mentally alike – as, for example, uniform rational actors – or as
overwhelmingly governed by forces that render distinctive worldviews
irrelevant – like class position or a universal set of material self-interests.
Second, ideological analysis reﬂects an awareness that we cannot simply
study the role of individual ‘ideas’ in isolation. To explain why human
beings buy into certain ideas, and to explain how and why those
ideas affect their behaviour in certain ways, we have to appreciate
how those ideas operate as part of broader systems of ideas. Taken
on its own, for example, the way many of the materially worst-off mem-
bers of liberal democracies support low spending on the sorts of social
services they would beneﬁt from, and simultaneously support tax cuts for
the wealthiest, looks almost inexplicable (Jost and Hunyady 2005). Yet
Ideological Analysis 299

such a view is often just one component of a set of ideological claims

regarding the historically demonstrated superiority of the small free-
market state, the need to minimize taxes and state spending in order to
promote (wealthy) ‘job creators’, the alluring promise that wealth and
success are accessible to every individual who works hard and the demand
to avoid dangerous ‘socialist’ notions that would create a slippery slope to
an authoritarian, unfree society. Once this interlocking system of ideas is
brought into focus, the reasons some individuals might accept such ideas
become clearer.
Ideological analysis is thus concerned with the excavation and forensic
examination of distinctive systems of ideas and the powerful role they play
in political life. Yet ideological analysis has frequently been undermined
by the infamous diversity of meanings that have been attached to the term
‘ideology’ (Eagleton 1991: 1–2; McLellan 1995: 1; Freeden 1996: 13, 47;
van Dijk 1998: 1; Humphrey 2005: 225, 227). To remedy this problem,
I offer a definition of ideology that reflects dominant contemporary
usages amongst theorists of ideology – which I suggest have finally started
to converge on a shared understanding (Leader Maynard and
Mildenberger 2016). But regardless of whether readers find my definition
amenable or not, it is always critical to clarify what one means by
‘ideology’. And in doing so, scholars must avoid what Matthew
Humphrey (2005: 299) has aptly labelled the stipulative error: justifying
particular definitions (against competitors) by superficially empirical
assertions about ‘what ideology is’, when how to classify different bits of
the empirical world is precisely what is being disagreed over in conceptual
disputes. Rather, different definitions should be justified according
to how functionally useful they are for research.
This functional usefulness should be assessed with reference to three
main considerations:
(a) essential features about the way ideology has been used in general
academic and lay discourse – for example, it would be a drawback of
a particular conception if it ended up asserting that phenomena
universally seen as ideologies, like liberalism, communism or social-
ism, were in fact not ideologies;
(b) the functional usefulness of the conception for the specific research
project at hand;
(c) the functional usefulness of the conception for broader academic
understanding across projects and disciplines – highly idiosyncratic
conceptions, that are liable to promote confusion or fragmenta-
tion in usage across different research communities, are
undesirable.
300 Chapter 14

Since (b) and (c) can pull us in opposite directions, striking a balance
between them is necessary.
So there needn’t be one single definition of ideology that all theorists
should use. But in this chapter I follow recent trends in ideological
analysis by suggesting that political theorists should use
a conception of ideology that is both broad and non-pejorative –
respectively, that it encompasses a large range of idea systems and
thought practices, and that it does not assume that ideology is necessarily
false, inflexible, poorly formulated or otherwise ‘bad’. Such a conception
is recommended by two of the most sustained conceptual investigations
into ideology, by Malcolm Hamilton (1987) and John Gerring (1997),
and has been favoured by many other leading theorists (Freeden 1996;
van Dijk 1998: 11–12; Jost 2006; Knight 2006). Following their advice,
I advocate the use of the following definition (closely resembling
Hamilton’s):
A political ideology is a distinctive overarching system of normative and/or reputedly
factual ideas, typically shared by members of groups or organizations, which shapes
their understandings of their political world and guides their political behaviour.

This definition is not so broad that all sets of ideas become indistinguish-
ably ideological, but it rejects assertions (by politicians or political
theorists) that certain political worldviews are ‘beyond ideology’ or ‘not
ideological but pragmatic’ (see also Bell 1960; Freeden 2005a; Coote
2014). Such claims frequently suggest exceptionalism, cryptonormativity
and an attempt at partisan and contestable political tactics rather than
sound conceptual distinctions (Žižek 1994b; Freeden 2005a; Worsnip
2015). In this broad and non-pejorative conception, ideologies denote
whatever distinctive idea systems people do in fact use to think about
politics. No human being can engage in some kind of perfectly rational
disembodied reflection about politics that simply ‘sees the world as it is’
uninfluenced by prior thinking. Instead, every individual’s political think-
ing occurs via networks of values, meanings, narratives, theories, assump-
tions, concepts, expectations, exemplars, past experiences, images,
stereotypes and beliefs about matters of fact already existing in their
mind (Geertz 1964; Berger and Luckmann 1967; Tversky and
Kahneman 1974; Wittgenstein 2001; Baurmann 2007). These networks
of ideas vary, at least somewhat, from person to person, group to group
and society to society, which is what makes them important objects of
study. In this conception, ideology is not reprehensible, but inescapable.
As Aletta Norval (2000: 316) writes: ‘It is this emphasis on the ubiquity of
ideology . . . that is at the heart of contemporary approaches to the
question of ideology.’
Ideological Analysis 301

This broad and non-pejorative conception stands in contrast to more

‘negative’ or ‘critical’ conceptions that have been historically prominent
in the study of ideology, especially in Marxist theory, but are increasingly
unpopular (van Dijk 1998: 3; Leader Maynard and Mildenberger 2016).
Such negative conceptions vary in the pejorative connotations they attach
to ideology – ideologies might be defined as ‘false consciousness’, or
oppression-legitimating systems, or especially fanatical or dogmatic
forms of belief, or some other unfavourable mode of thought.
Advocates of such conceptions of ideology frequently argued that such
connotations are essential for the concept of ideology to retain any critical
power and normative relevance (Larrain 1979: 15, 52, 118; Thompson
1984: 4, 12, 82; Boudon 1989: 24–30; Rosen 2000: 393–5). But this
argument has consistently been offered as a non-sequitur. There is no
reason why a concept needs to be defined pejoratively in order to be useful
for critical or otherwise normatively evaluative work (Larrain 1979: 77;
McLellan 1995: 23; Freeden 2005a: 262; Steger 2008: 4–5). As Teun van
Dijk (1998: 11) argues
Does [the] more general conception of ideology take away the critical edge of the
enterprise, as is sometimes suggested, or prevent ideological critique? Of course it
does not. No more than that the use of the general concept of ‘power’ precludes
a critical analysis of power abuse, as well as solidarity with the forms of counter-
power we call resistance.
What is lost with a broad and non-pejorative conception of ideology is the
ability to critique something simply by labelling it ideology. But this sort of
terminological smearing is not the limit of normative critique, nor
a persuasive form of it. And a key problem with pejorative conceptions
of ideology is that they tend to encourage a prejudicial form of analysis.
It is notable that some of the major non-Marxist figures to deploy them
were, ironically, anti-Communist scholars during the Cold War. For
these figures, the concept of ideology denoted ‘extremist’, ‘totalitarian’
and ‘anti-modern’ views outside the mainstream of Western politics.
Such an understanding amounted, as Terry Eagleton (1991: 4) aptly
caricatures, to the claim ‘that the Soviet Union is in the grip of ideology
while the United States sees things as they really are’. In these examples
and others, pejorative definitions conceptually encode the sorts of
normative commitments political theory is intended to render explicit
and subject to interrogation and reflection.
By contrast, broad and non-pejorative conceptions allow us to conduct
a more open-minded and rigorous form of analysis where, as Humphrey
(2005: 237) puts it, ‘ideological forms are not presupposed but emerge
through careful empirical analysis of thought instantiations.’ And they
302 Chapter 14

also allow us to avoid high levels of fragmentation in scholarly under-

standings of ideology, by keeping the master concept inclusive, but
leaving room for the identification of subtypes of ideology that might be
dogmatic, irrational and so forth.
We should therefore maintain a distinction drawn by Freeden (1996:
27–8, 133; 2013b: 52–3) between ideological analysis (the study of real-
world ideologies) and ideologizing (the construction of ideologies, which
occurs partly through political theory). These two activities are always
entangled. Since we all think about politics under the influence of our
various respective ideologies, there may be no completely neutral social
science or analytical philosophy. But it is vital to recognize a difference
between the construction of our own concepts and normative principles,
and the efforts to find out what are or were the extant concepts and
normative principles (and other ideas) of others (Weber 2009: 145–6).
This chapter explains how to do the latter in ways that inform, support or
otherwise take a central role in the former.
I make three further remarks about the meaning of ideology. First,
an important distinction should be drawn between what we
might call personal ideologies, by which I mean the particular
ideologies of individual people, and shared ideologies, that
describe the systems of ideas held in common by groups.
Ideologies ultimately exist in minds (though they have emergent social
aspects too), and groups do not truly have minds, so imputing ideol-
ogies to groups represents something of a metaphorical abstraction,
though a benign and often productive one (Thagard 2010). Every
member of a group inevitably has slightly distinct ways of thinking –
slightly different personal ideologies. But scholars can and usually do
talk of shared ideologies to make generalizations about important
similarities between the personal ideologies of individuals, and to
draw attention to emergent properties of ideologies as they become
consciously identified in social discourse and embedded in institutions
and practices. There is nothing mysterious about this in and of itself
(though, of course, such generalizations can be formulated erro-
neously), and a productive analogy can be drawn with the way we
talk about languages. Just as every ‘conservative’ thinks in a somewhat
different way, so every speaker of ‘Spanish’ or ‘Urdu’ will speak in
a unique, idiosyncratic manner. But in both scholarly and lay talk, we
can productively talk about the language they speak in common, in
a way that generalizes about key similarities without denying indivi-
dual variation (van Dijk 1998: 30). So too with ideology. And like
languages, the ideologies we talk about can involve varying levels of
generalization – from low (Al Qaeda leadership ideology, IMF
Ideological Analysis 303

neoliberalism) to medium (Rawlsian liberalism, Stalinism) to high

(socialism, conservatism). Labelling ideological subtypes in this
way is useful for retaining conceptual precision in the sorts of
ideologies we are talking about.
Second, it is important to appreciate that ideologies are sub-
stantively rich phenomena. It is common to talk about ideologies as
deﬁned by certain core concepts, values, political ambitions or domi-
nant narratives. But ideologies are built from all of these and a vast
array of other sorts of idea or idea cluster: identities, myths, memories,
stereotypes, epistemic rules, beliefs about matters of fact, rhetorical
repertoires, strategic preferences, exemplars, expectations, horizons of
possibility, images, lived experiences and so forth. Human beings
differ in their thinking on all of these, often in important ways that
exert a powerful effect over their broader thinking and behaviour.
By thinking of ideologies as complex networks of a vast array of
notions, one can engage seriously with the idiosyncratic forms of
thinking that characterize particular individuals and groups in the
real world.
Third, while I advocate a somewhat ‘cognitive’ understanding of
ideologies, it is vital to recognize the inextricable relationship
between ideology and discourse (van Dijk 2013). Since it is arduous
and ineffective to invent new ways of talking from scratch on every
issue, ideologies are characterized by certain ways of talking, certain
rhetorical repertoires and certain arguments and justiﬁcations. These
are picked up and reused by individuals to engage in argument, legit-
imation, persuasion, sophistry and so forth. And individuals’ ideologies
are, in turn, shaped by the discourse of themselves and others – as new
ideas are encountered in communication, engaged with and rejected or
internalized. Many ideologies are also represented socially in discourse,
picking up social meanings and connotations, and they may inhabit
social movements and become embedded in social institutions and
groups through discourses and practices (van Dijk 1998: ch. 3).
Ideological analysis is thus inevitably concerned with not just how
people think, but also how they talk and act.

3 Why Should Normative Political Theorists

Use Ideological Analysis?
In this section I provide four main reasons to think that ideological
analysis can be of deep relevance to the kind of normative political theory
that is a key focus of this volume.
304 Chapter 14

3.1 Evaluating Institutions and Ideologies

Perhaps most obviously, ideologies may be essential components of the
operation or legitimation of various institutions and social processes that
are problematic (or progressive) from a normative political theory stand-
point. Put simply, ideologies have major social and political effects, and
are involved in the use of social and political power. There are two key
ways in which ideological analysis supports political theory in grappling
with such effects.
First, ideological analysis allows us to diagnose the normative
failings of existing political institutions by illustrating the proble-
matic ideologies they generate (or, conversely, identify normatively
beneficial aspects of institutions by showing the beneficial ideologies they
generate). For example, diagnoses of the normative dangers of unregu-
lated private media in emerging democracies should examine the way in
which such media tend to produce nationalistic and racist political dis-
courses (Price 2000; Mann 2005). Similarly, assessments of the value of
two-party over multi-party electoral systems should consider how the
former may narrow the ideological landscape of society (which could be
seen as good or bad), encouraging people to coalesce into just two major
ideological camps, but also incentivizing movement towards ‘centrist’
and away from ‘extremist’ ideological positions (Sartori 1976: 178–9;
Evans 2002). Analysing how such outcomes are created, why they
might be normatively problematic or desirable, and the normative per-
missibility (and efficacy) of potential political responses, will all rest on
ideological analysis.
Second, ideological analysis allows us to critique particular
ideologies by illustrating the flawed socio-political institutions
they generate or sustain (or, conversely, identify strengths of particular
ideologies by highlighting their role in desirable socio-political arrange-
ments). For example, free-market ideologies may reproduce idealized
understandings of property and markets that efface awareness of racial
hierarchies in the real-world economic system (Mills 1999; Shelby 2003).
But this suggestion relies on empirical claims about the impact certain
ideological notions have on political thinking and, thereby, on economic
and political behaviour. Such ideological analysis may include interroga-
tion and critique of certain concepts or ideas used by political thinkers and
actors (see Olsthoorn’s chapter in this volume). This may reveal histori-
cally changing connotations and assumptions that are involved in
particular speech acts, and shed light on the impact of different inter-
pretations of different concepts on political outcomes (Ball et al. 1989: ix;
Freeden 1996: 100–17; Skinner 2002b: 114–21, 158–74). Again, all such
Ideological Analysis 305

critique of existing systems of political thought and talk needs to be

grounded in ideological analysis.
Frequently, the ideological notions that produce institutions or prac-
tices, and the ideological notions those institutions or practices produce,
are simultaneously relevant. For example, evaluations of the policy of
racial proﬁling might be concerned with the way its political rationale
relies on crude racial constructs and ignorance of structural causes of
crimes, and with how it encourages racist stereotypes of certain ethnic
groups as having a propensity to criminality (Shelby 2003: 175–6).
Evaluations of fee-paying schools might be concerned both with the
way they entrench misconceptions of the added educational value such
schools generate (if much of their superior performance is simply because
they attract already high-performing students) and with how such mis-
conceptions legitimate and sustain support for private schools (Swift
2003: 21–3). Most obviously, Marxist and post-Marxist traditions of
normative theory have long engaged in analysis of both how ideology
serves to legitimate domination, political exclusion and various other
forms of political injustice and the way those practices reproduce the
legitimating ideology (Žižek 1994b; Laclau 2007). Given the vital role
ideology continues to play in the legitimation and operation of a panoply
of leading political and social institutions, this form of ideological analysis
(shorn of implications that ideology is always malign) ought to feature
more extensively across contemporary normative political theory.
More generally, normative evaluation of the general role of ideologies in
political life is central to many questions in political theory. Ideological
diversity is a critical ingredient of a free society, and the fact that human
political thought and behaviour occurs under the inﬂuence of distinctive
ideologies is non-contingent and inescapable. All political theory – ideal and
non-ideal alike – therefore needs to grapple with this general role of
ideology. The ideological nature of human beings may pose obstacles to
certain accounts of rational citizenship, deliberation in democratic dis-
course, solidarity in a civic community, neutrality in state institutions and
fair distribution of political, communicative, epistemic and material
resources. Ideological analysis is therefore relevant to the comparative
normative evaluation of theories and institutions with respect to all these
questions.

3.2 Assessing Political Theory’s Principles in Real-World Contexts

The ﬁrst reason for thinking that ideological analysis is relevant to
political theory highlights the relevance of ideology to all sorts of norma-
tive political theorizing. But there is a further reason to think ideological
306 Chapter 14

analysis important to the degree that political theory moves away from
ideal-type theorizing and becomes ‘non-ideal’ (Valentini 2012) or ‘realist’
(Galston 2010). In such situations, ideological analysis is needed to
assess the likely real-world effects of certain moral or political
principles in practice, and consequently their viability and normative
attractiveness. Such forms of non-ideal normative political theory are
concerned not just with reflection on moral rightness in some ‘ultimate’
sense, but the actual application of political theory to the world, and the
design of political prescriptions and norms to guide policy. How real-
world agents – perceiving, thinking and acting under the influence of real-
world ideologies – would actually operationalize normative principles is
thus a key question. As Jennifer Welsh (2010: 424) notes: ‘principles,
when adopted, always take on a life of their own. It is difficult to control
their meaning, or to avoid their misuse.’ Predicting how this ‘life of their
own’ will unfold ought to be a critical component of the assessment of
normative principles at the non-ideal level.
Ideological analysis supports these sorts of non-ideal and realist assess-
ments in four ways. First, it may provide contextual knowledge about
actually existing ideologies in particular circumstances. Failure to analyse
the nature and power of these extant ideologies can undermine normative
theorizing about the right way to approach injustice. For example, a range
of left-wing terrorist groups in the 1970s and 1980s believed that a few
symbolic acts of violence would expose the fragility of the capitalist order
and inspire a mass revolution, and acted accordingly (Tsintsadze-Maass
and Maass 2014). This legitimation of deadly political violence involved
a wildly naïve failure to appreciate the power and breadth and depth of
internalization of liberal, free-market ideologies in modern Western
societies. A parallel problem might be thought to exist in neoconservative
or neoliberal theories of why military intervention in countries to spread
democracy should be legitimate. Such theories carry many flaws, but
these include a fundamental blindness to the complex ideological terrain
of target states, and the radicalizing ideological consequences of militar-
ized action on both its victims and on the interveners themselves (Doris
and Murphy 2007; McCauley and Moskalenko 2011). The problem is no
less salient for other political strategies – one needs to know the ideologi-
cal context of a proposed policy to predict the normatively relevant
consequences. Normative political theorists have not been completely
blind to this. Members of the International Commission on
Intervention and State Sovereignty (ICISS), for example, advanced the
international norm of a ‘responsibility to protect’ in 2001 partly in the
belief that this concept would do a better job than the previous concept of
‘humanitarian intervention’ at reconciling competing attitudes and
Ideological Analysis 307

commitments in the extant ideologies of Western democracies,

developing world states and authoritarian great powers (Evans 2011).
This, it was hoped, would be more efficacious in giving ‘the responsibility
to protect’ the force and political traction needed to save lives.
Second, ideological analysis may reveal general tendencies in political
thinking – whether rooted in basic psychological propensities with roots
in human nature, cultural dispositions or something else – which are
critical in foreseeing how political ideas will play out in practice. For
example, leading political psychology research on ideology demonstrates
strong human tendencies towards ‘system justification’ and ‘just-world
thinking’ (a predisposition to rationalize the existing political system and
its consequences as just). This research also shows that humans have
a preference for epistemic satisfaction, i.e. achieving a set of beliefs that
seem satisfying for a range of underlying psychological motives, rather
than epistemic optimization, i.e. making sure that one is actually right
(Furnham 2003; Jost and Hunyady 2005). This is highly relevant for non-
ideal political theory, since it problematizes any assumptions that most
individuals will respond to policies as rational, contemplative actors, and
reveals key psychological supports for injustice and barriers to political
and social reform. Again, political thinkers have not completely eschewed
this line of theorizing. John Stuart Mill’s (2008) defence of free speech
rested fundamentally on empirical assumptions about how political
thinking evolved, progressively, in a society, and about how false or
oppressive ideological notions could best be countered given the way
human thinking and discourse worked. But the validity of such an
account rests on claims about these real-world dynamics of ideological
change – necessitating ideological analysis.
Third, partly but not solely through the first two ways, ideological
analysis may provide the methods and skills for reflecting on how a certain
normative system or prescription will play out in the political thinking of real-
world actors – focusing not on logical implications of arguments and claims
under rigorous philosophical analysis, but the likely forms of reasoning,
assumptions and attitudes such arguments and claims might encourage in
actual political practice by citizens and elites. Once again, examples of
such an approach already exist. Isaiah Berlin’s famous (1969) critique of
‘positive liberty’ did not necessarily deny any normative attractiveness to
that notion in the abstract. Many of Berlin’s concerns lay instead with the
attitudes and assumptions positive liberty subtly encouraged, and the
sorts of pernicious political programmes it could be mobilized to legit-
imate. In the sense I mean it here, Berlin’s political theory involved a form
of ideological analysis: imagining how a concept would operate in ‘actual
political thinking’ (Freeden 2008: 197).
308 Chapter 14

Fourth and ﬁnally, a range of context-speciﬁc questions for political theory

may also arise at the non-ideal/realist level for which ideological analysis is
particularly relevant (see also the chapters by Jubb and by Schmidtz in
this volume). How, for example, should we know what political institu-
tions or policies to implement in a state formerly under an ideological
monopoly but now emerging into competitive party politics? Or in
a society increasingly polarized across apparently irreconcilable ideologi-
cal divisions? Or in a state whose unity and future stands in question in the
face of nationalistic ideological groups? Or in a context where radicalizing
violent ideologies appear to thrive through institutions long thought
essential to free speech? To grapple with such questions in detail, political
theorists need a developed understanding of how ideologies operate in
these various contexts.
In all such instances, particular normative concepts or principles are
being advanced in non-ideal theory because of assessments of extant
ideologies, the ideological tendencies of human beings in general, how
particular normative claims would work in practice as ideology or the
specific ideological dynamics of real-world problems. Yet whilst I have
noted that such lines of theorizing are more common than is perhaps
consciously appreciated, they are often grounded on speculation by poli-
tical theorists rather than rigorous, empirically grounded ideological
analysis. Consider, for example, Raymond Geuss’s (2003: 285–6) inter-
esting but entirely unsubstantiated claim that Rawlsian political theory
underpinned neoliberal economic policies and a rise in global inequality
in the 1970s and 1980s. This could be investigated through ideological
analysis. Instead, Geuss bases his argument on the mere correlation of
Rawls being influential in American universities with rising inequality –
a classically invalid causal inference (see also Sagar 2014). Even ICISS,
Mill and Berlin, though they do rest their arguments on some sort of
ideological analysis, do not deploy any rigorous or developed methodol-
ogy for doing this. Consequently their conclusions are contentious: it is
not clear that R2P has successfully evaded the ideological connotations of
humanitarian intervention (Chandler 2004), substantial evidence exists
against Mill’s theories of the epistemic benefits of free speech (Edelman
1977; Vigna and Gentzkow 2010; Ipsos MORI 2014), and it is question-
able whether Berlin’s distinction between negative and positive liberty
effectively maps onto real-world divergences in actual political thinking
(see also Skinner 2002a: 238–9; Nelson 2005). More generally, norma-
tive political theorists interested in non-ideal or realist approaches have
not engaged in detail with ideological analysis (Freeden 2012).
So, done in a sophisticated manner, ideological analysis could increase
the solidity of the empirical assumptions on which non-ideal and realist
Ideological Analysis 309

political theory arguments rest. It could even provide empirical evidence

of how particular normative concepts or principles have played out in
practice in the past. Ideological analysis might reveal, for example, how
notions of liberal toleration can encourage unintended belief in moral
relativism, how claims about inheritance rights can undermine awareness
of inequality in the initial distribution of resources or how arguments that
the targeting of ‘human shields’ is permissible can encourage a cognitive
slippery slope in the perception of when violence against civilians is
legitimate. None of these effects deﬁnitively establishes the unacceptabil-
ity of the normative beliefs or claims in question. But, if they exist, they
may be considered relevant marks against such beliefs or claims. This
form of ideological analysis of normative concepts and principles might
be particularly crucial to any form of ‘critical’ political theory (in the
broadest sense) that seeks to expose the unintended justiﬁcatory power
of various normative concepts and principles.

3.3 Exposing Ideological Assumptions behind Arguments in Political

Theory
Ideological analysis also allows an interrogation of the validity of
normative arguments in and of themselves by exposing their
background ideological assumptions. Sophisticated ideological
analysis involves seeking out the full range of contextual, idiosyncratic
notions within which individual claims are embedded, which shapes their
speciﬁc meaning, and which explains their plausibility to those espousing
them. Normative political theorists are very used to seeking out the
principles and values that logically precede certain claims, but these
may not be the only sorts of ideas or the only form of support that
ideological notions lend to an argument. Revealing the beliefs about
matters of fact, the narratives, the familiar stereotypes, the particular
interpretations of concepts and so forth connected to a certain argument
may be vital for both (a) understanding why another political theorist
(including a historical thinker) considers that argument plausible, and (b)
thoroughly assessing the argument’s defensibility.
Perhaps the most important use of this mode of ideological analysis is
that which serves to maximize a theorist’s own self-critical rigour – sub-
jecting their own political theory to ideological analysis, to unpack not just
the analytical assumptions on which the argument formally rests, but the
other attached notions that might explain the moral intuitions behind key
claims. Ideological analysis might, for example, expose particularistic
commitments rooted in certain moral worldviews, cultural perspectives
or historical epochs that are wrongly presumed to reﬂect universal moral
310 Chapter 14

assumptions (which is not to presume that there can never be such

universals, just that particularistic beliefs can often masquerade as them).
Political theorists and philosophers are sophisticated thinkers, but not
without their own ideological environments and standpoints, nor are they
immune to the well-researched psychological tendencies like just-world
thinking or consistency bias mentioned earlier. Rigorous political theory
should critically examine those environments, standpoints and possible
tendencies in analysing claims by political theorists themselves.

3.4 Encouraging Creative Political Theory

Finally, ideological analysis can be a powerful method for spurring crea-
tivity in formulating new ideas, concepts and arguments. I gently suggest
that many sections of normative political theory – whilst valuable and
sophisticated – are not as innovative as they could be, taking the form of
seemingly interminable debates over a narrow set of basic questions
between a narrow set of well-established standpoints. Ideological analy-
sis should be used to enable a rigorous engagement with the
thought systems of other periods in history, and other cultures in
our present (see the chapter by Ackerly and Bajpai in this volume).
In doing this we might discover new concepts, ideas and claims that
open up unexplored political-theoretic spaces (though it is important to
avoid a ‘touristic’ approach, where ideas seem alluringly exotic but are used
anachronistically and without real understanding). When one considers all
the possible (and historically existing) ideological positions that can be
generated, it should be clear that contemporary political theory in predo-
minantly Western academic institutions occupies only a small portion of
the ideological universe of possibilities. Many of the potential normative
positions may be unexplored for good reason – being patently incoherent,
radically counterintuitive or reprehensible in light of moral commitments
that we are loath to part from. But it is unlikely that contemporary political
theory exhausts the plausible normative realm. The classic example of
a form of such ideological analysis yielding notions of normative interest
concerns ‘republican liberty’, which has frequently – from the nineteenth-
century writings of Benjamin Constant to the contemporary work of
Quentin Skinner – been rooted in an ideological excavation of earlier
systems of political thought (Constant 1988; Skinner 1998). And ideolo-
gical analysis may not only support such projects through the investigation
of existing but temporally or culturally distant ideologies. It may also
support innovative normative political theory by allowing us to schematize
and map out potential normative positions, irrespective of whether these
Ideological Analysis 311

have been actually espoused by historical or contemporary cultures

(Freeden 2013a: 118).

3.5 Summary: The Need for Ideological Analysis in Normative Political

Theory
These four uses of ideological analysis suggest that scholars doing
normative political theory should take ideological analysis
seriously. This point needs wider recognition, since all four will be
increasingly important enterprises as the discipline looks ahead to the
rest of the twenty-first century, for two reasons. First, many of these uses
of ideological analysis for normative political theory assume that one is
engaging in a fairly non-ideal, empirically engaged and perhaps applied
project, or at least serve that sort of political theory in more obvious ways
than they serve ideal theory. Such forms of non-ideal political theory are
increasingly popular, creating an increasing need for rigorous ideological
analysis. This should not, though, be taken to marginalize ideal theory or
obscure how my first, third and fourth uses for ideological analysis all
could (and should) be deployed in ideal political theory.
Second, in the remainder of this century, political theory is inevitably
going to become more internationalized and globally integrated, with deeper
engagement between Western intellectual traditions and those from the
many other parts of humanity’s cultural heritage. This is going to be
a challenging but intellectually vital process. Cross-cultural fertilization in
political theory will expose divergences on issues and assumptions long
thought settled within particular academies and paradigms. And it will
force scholars from all nations to encounter claims, arguments, theories
and political problems that they struggle to understand on a superficial
examination, rooted as they are in assumptions, beliefs, ways of reasoning
and contexts unfamiliar to the scholars’ own ideological positions. For such
engagement to be productive rather than lead to radical misunderstanding
and bemusement, political theorists on both sides of cultural, ideological
and theoretical divides are going to have to get better at ideological analysis,
especially in my second and third senses: to properly interpret, interrogate
and engage with each other’s positions, and evaluate the (possibly varying)
applicability of certain normative concepts and principles to diverse ideolo-
gical contexts.

4 How to Do Ideological Analysis

As I noted in my introduction, a problematic feature of the developing
ﬁeld of ideological analysis is that work on ideology remains very
312 Chapter 14

fragmented across disciplines. Since these different disciplines generally

display contrasting strengths and weaknesses, I suggest that if we are to
get to grips with ideology, it is vital to deploy an integrative approach
(the best cross-disciplinary compendium is provided by Freeden et al.
2013). Intellectual historians and political theorists have generated
sophisticated methods for examining and conceptualizing the content of
ideology. Political psychologists and sociologists have conducted exten-
sive research on some of the causes of ideological attachment – why people
internalize the ideological positions they do. Political scientists and dis-
course theorists, meanwhile, have studied extensively the effects of ideol-
ogy on behaviour. Yet all three of these aspects of ideology are relevant for
political theorists.

4.1 Uncovering Ideological Content

Studying ideology is fundamentally about examining how actual people
think. As such, to say anything of consequence about ideology, we need
to find out something about the actual contents of the ideology in
question. Too often, scholars studying ideology eschew this task. Instead
they presume that ideological content is familiar and uncomplicated, by
reducing it to a label, or a point on a crude dimension – such as the seven-
point scale from ‘Extremely Conservative’ to ‘Extremely Liberal’ used in
US election surveys – which is often presumed to self-evidently denote
a set of policy preferences, or normative standpoints, or value attitudes.
This might be defensible for certain limited forms of correlational analysis
in political science, but it is of little use to normative political theory (and
most other fields). Consequently, effective ideological analysis
requires that researchers deploy methods that actually collect
rich data on, and engage seriously with, real-world political
thinking (Freeden 2013b).
I suggest that there are four primary methods of collecting such data: (i)
behavioural inference – reaching conclusions about individuals’ ideologies
in light of the ways they behave; (ii) textual analysis – examining ideas
expressed in any forms of communication, including non-verbal works,
which exist independently of the analyst; (iii) inquiry – the attempt to
directly solicit ideas or beliefs out of individuals through questioning –
including qualitative methods like interviewing and quantitative methods
like surveying; and (iv) neuroscientific methods – the study of neurological
processes in an individual’s brain and wider nervous system. I will con-
centrate on inquiry and textual analysis, the two methods likely to be most
useful for political theorists interested in real-world ideologies.
Ideological Analysis 313

Defensible ideological analysis rests on interpreting texts and

responses to inquiries rigorously and accurately – drawing plausible
conclusions about individuals’ thinking from the data contained
therein, including non-verbal instances of communication. The deep
methodology of this practice is the subject of an extensive literature
and considerable debate in the discipline of hermeneutics and broader
study of interpretation. It is beyond my capacity to reproduce that
literature here (but see Blau’s chapter in this volume, as well as
Skinner 2002b; Yanow and Schwartz-Shea 2006; Wodak and Meyer
2009; Fairclough 2010). Instead I wish to list some key practical
pointers on how to interpret data in order to build pictures of the
ideologies of certain individuals or groups.
The most important pointer in this respect is that to generate reliable
accounts of an ideology, researchers need to avoid the cardinal
hermeneutic sin of acontextualism. I use this label to refer to the
general tendency to assume that individuals, across cultures, spaces and
times, generally hold a set of basic perceptions, beliefs, values and
interests pretty close to that of the analyst, or the analyst’s society
(Boudon 1989: 74–80, 95–6; Skinner 2002b). The contrary fact that
individuals actually hold highly diverse sets of such beliefs doesn’t imply
that individuals are wildly irrational, or deny that there are also com-
monalities in the way individuals think. But as Skinner points out: ‘what
is rational to believe depends in large measure on the nature of our other
beliefs’ and as a result, we should ‘interpret speciﬁc beliefs by placing
them in the context of other beliefs [and] interpret systems of belief by
placing them in wider intellectual frameworks’ (Skinner 2002b: 4–5).
This includes asking what it is feasible to assume individuals were
expressing and thinking given the intellectual resources and dominant
ideological environment in which they were writing/speaking (Skinner
1974: 283, 289; Baurmann 2007).
Matters are further complicated, however, by the possibility that
individuals may be engaging in dissimulation, irony, rhetoric or some
other form of speech that does not sincerely reﬂect their views. Hence the
importance of attending to not just what bearers of ideology are saying but
what they appear to be doing – e.g. lobbying for support, trying to per-
suade, satirizing another, making the interviewer happy, saying what they
‘ought’ to, etc. (Tully 1983: 490–4; Skinner 2002b: 2–4; George and
Bennett 2005: 99–108). Ideological analysts must attend to the pos-
sible strategic purposes of a particular text or expression and to
the intended audience.
The obvious implication of contextualism is that ideological analysis
rests in large part on gathering considerable information on the context of
314 Chapter 14

the ideology in question. This might be done through examining relevant

historical or cultural studies, through absorbing as much of the discourse
produced by the individual or group we are studying as possible, through
studying primary data (i.e. other cultural products and discourse frag-
ments) from the relevant context and so forth. The burden here could
easily become unmanageable for a normative political theorist who is not
primarily an ideological analyst, but interested in ideological analysis for
the four reasons I listed in Section 2. But a more modest contextualism is
possible, where we take seriously underlying differences in the
cultural and political environment of those we are studying and
ourselves, and are aware of the necessarily tentative nature of
interpretations of the thinking of others in light of uncertainty
about their broader ideological assumptions derived from their
social context.
With such a hermeneutic sensitivity, researchers might proceed by
asking a series of structured questions about the ideological data
(texts or interview responses) they are able to gather. Exactly what
questions researchers ask will relate to their objectives, but I shall propose
a possible core list as a starting point. The speakers (or authors) being
studied might be institutions, movements, political parties or other group
entities, as well as specific individuals, but we should still ask who are the
actual specific individuals who have authored apparent
statements of group entities as a whole, and how ideologically
heterogeneous might the broader group be.
1) What has the speaker seemed to say, and what does he or she seem to
be doing in saying this?
2) What are the concepts and terms that make up the speaker’s expres-
sions, and what are the most plausible meanings of those concepts and
terms the speaker would attach to them? (In interview methods, such
meanings might be solicited by further questioning.)
3) What beliefs, arguments or other ideas do these concepts and terms
together seem to express?
4) Are there reasons to believe that the speaker does not sincerely hold
those beliefs/arguments or other ideas? On all available evidence, does
it seem fair (it will never be certain) to assume that these reflect the
speaker’s own views? If not, why are they being said – and what might
this tell us, if not about the speaker’s own ideology, about the broader
ideologies of others in the context in which he or she is speaking?
5) What possible ideas, not explicitly expressed by the speaker, (a) might
lend comprehensibility and consonance to his or her overall ideologi-
cal system and (b) might there be good reasons to believe the speaker
held consciously or unconsciously?
Ideological Analysis 315

6) What appears to be the ideological structure of the speaker’s ideas?

By this I mean the relationships between the various ideas – relation-
ships that might be logical (presupposition, inference, entailment,
etc.), semantic (connotation, inclusion, exclusion, clarification, is
a subset of), emotional (attraction, antipathy, felt association), cau-
sal (producing, requiring, precluding, encouraging) or something
else?
By asking these questions, a researcher can start to put together a picture
of the system of ideas that constitutes an individual’s ideology, or the part
of his or her ideology relating to some particular defined question or issue.
That picture might be rendered in detailed, prose description, it might be
conceptualized dimensionally on a single dimension or multidimensional
space (although, as I have suggested, this may oversimplify ideological
content too heavily for the political theorists) or it might be visualized
through a form of concept mapping, such as the Cognitive-Affective
Mapping methodology developed by Paul Thagard and colleagues at
the University of Waterloo (Thagard 2012). The reliability of this picture
rests on the open-mindedness, free exploration and interpretive honesty
of the researcher. It is all too easy for scholars, superficially deploying
appropriate data collection methods, considering context, and asking the
right sorts of interpretive questions, to ‘see what they want to see’.
Ideological analysis must always attend to this danger of contrived,
biased analysis, explicitly consider alternative interpretations and
explain why they have been rejected.

4.2 Explaining Ideological Attachment

I already mentioned the strong tendency for individuals to engage in
system justiﬁcation and just-world thinking: to adopt the default assump-
tion that outcomes are fair, with individuals deserving what they get, even
when evidence for this is lacking (Jost et al. 2003; Jost and Hunyady
2005). That this is often a default assumption obviously doesn’t mean it
is always felt – perceptions of unfairness and injustice are fundamental
and widespread in human thinking. But it is just one of the pieces of
evidence that shows how ideological belief formation is driven by a broad
range of underlying psychological motives and social processes (both of which
are deeply interrelated). Political theorists may frequently be concerned
with the role of such psychological and social causes of ideology in their
analysis.
Key psychological motives driving belief formation include:
(i) epistemic motives – sincere concerns with working out what is true;
316 Chapter 14

(ii) cognitive dissonance minimization (also known as consistency bias) –

a desire to maintain consonance between one’s beliefs, and beliefs
and actions, even when this involves adopting beliefs not supported
by evidence;
(iii) self-esteem motives – the need for a sense of self-worth, social standing,
superiority, belonging, recognition and life purpose;
(iv) cognitive efficiency concerns – the drive to avoid overly burdensome
mental activities;
(v) comprehensibility motives – the need to find the world comprehensible,
determinate and not worryingly uncertain;
(vi) anxiety suppression – for example, about death, ethical propriety,
sexual status, the wrongness of past decisions and commitments
and so forth.
This is not a comprehensive list but highlights some of the key findings in
psychological research that could be of relevance in considering how
political ideologies form (see, in general: Tversky and Kahneman 1974;
Boudon 1989; Pinker 1998; Boudon 1999; Jost and Major 2001; Jost
et al. 2009b; Kahneman 2012; Varki and Brower 2013).
Key social processes driving belief formation include:
(i) ideational resources – the availability of ideas, concepts or frameworks
of reasoning in the discourses accessible to the individual;
(ii) discursive saturation – the saturation of ideas into discourse within the
particular social networks or media environments in which the indivi-
dual is embedded, such that those ideas appear to be ‘common sense’;
(iii) epistemic dependence – where an individual relies on certain epistemic
authorities that appear credible (or that the individual lacks more
credible alternatives to) for information that is difficult for them to
personally verify;
(iv) groupthink – the tendency of individuals to adopt the beliefs, or at
least the avowed beliefs, of the majority around them;
(v) ideological incentivization – the attachment of desirable material or
symbolic outcomes to the adoption of a belief, or at least the avowed
adoption of a belief;
(vi) rhetorical presentation – the skilful deployment of rhetorical devices
and emotional appeals in communication that encourage internali-
zation of certain ideas.
Again, this list is a tentative summary of major themes in existing
research, rather than a comprehensive list (see in general: Berger and
Luckmann 1967; Hardwig 1985; Boudon 1989; Simonds 1989;
Fairclough 2001; Skinner 2002b; Baurmann 2007; Rydgren 2009).
These psychological and social processes that shape ideological attach-
ment should help political theorists construct plausible pictures of the
Ideological Analysis 317

ideologies of real-world individuals, explain a wide range of just and unjust

states of affairs, and assess the viability of certain political prescriptions and
strategies. They are at the root of why certain sorts of legitimating practices
are effective, certain key realities denied, certain human consequences of
actions or policies ignored. That is not to suggest that such processes are
immutable, universal or constant across individuals and societies – on the
contrary, they vary and might themselves be targets of political reform by
prescriptive theory. But this is just another reason to believe that political
theorists should attempt to understand and analyse them.

4.3 Analysing the Effects of Ideologies

But constructing an accurate understanding of extant ideologies, and the
psychological and social processes that encourage them, is unlikely to be
the only activity of interest to the political theorist. Ideological analysis
should then involve an examination of what political and social
effects these ideologies have.
Ideologies provide cognitively necessary repertoires for thinking and
socially comprehensible repertoires for talking: they enable, constrain
and shape political thought and discourse. They constitute individuals’
own internalized thinking about subjects and provide them with ideas,
arguments, concepts and claims to instrumentally use (for a range of
sincere or strategic purposes) in communication with others.
The broadest effect of ideology, via both these avenues, is to provide
individuals and groups with the intellectual resources used to think and
talk about that world, thereby constructing their understanding of social
reality (Searle 1995; Tileag ă 2007: 722; Thagard 2012: 51–2). But in
particular, ideologies shape individuals’ conceptions of what is desirable,
and what is permissible – as I term it, they motivate and legitimate certain
forms of action, including the establishment and maintenance of institu-
tions and other political arrangements.
Unpacking these effects comprehensively would take at least a book.
Aiming only to generate ideas and focusing attention for researchers,
I shall provide a fairly ad hoc list of particularly signiﬁcant sorts of beliefs
that are generated by and constitute ideologies. Ideologies contain:
(i) Notions of the proper members of a society or group, and conceptua-
lizations of the borders between those proper members and others,
between the in-group and the out-group. In particular, ideologies
shape notions of differential obligations owed to different categories of
person (Tajfel 1974; Opotow 1990; Huddy 2001; Tileagă 2007;
Hammack 2008).
318 Chapter 14

(ii) Latent assumptions or explicit judgements regarding the prototypi-

cal, expected and proper behaviour or characteristics of members of
a society or group. In particular, ideologies determine the charac-
teristics/attitudes/behaviour to which praise, glory and status, on
the one hand, and shame, disempowerment and exclusion, on the
other, are attached (Huddy 2001: 133–7, 143–5; Finlayson 2012).
(iii) Specification of key objectives, ends and anticipated futures towards
which a society or group is aiming, and programmes of collective
action for getting there (Mullins 1972: 510).
(iv) Assertion of values and objects, perhaps of a quasi-sacred or truly
sacred nature, to which individuals and groups feel deep emotional
attachment (Atran and Axelrod 2008).
(v) Identification of key problems, threats, challenges or obstacles to its
objectives and ends that a society or group faces (van Dijk
1998: 66).
(vi) Conceptions of the relevant ‘field of possibilities’ – the set of policies,
strategies or institutional arrangements that are feasible, imaginable
or otherwise warrant consideration. What is excluded from this
field of possibilities (presented as impossible, infeasible, futile,
idealistic, utopian, etc.) may be as important as what is included
(Foucault 1982: 789–92; Crawford 2002: 20).
(vii) Specific normative prohibitions and normative obligations, for indivi-
duals, for collective actors or institutions and for a society at large
(Crawford 2002; Van Dijk 2013: 74–7).
(viii) Characterizations, whether impressionistic or detailed, of rival groups
and competitors, within or without a polity, and their objectives,
characteristics, own ideologies and current activities (Shenhav
2006; Hammack 2008).
(ix) Particular beliefs about matters of fact, whether true or false, whether
inchoate and vague or specific and dense, whether rooted in narra-
tives and imagery or statistics and apparent evidence (Hochschild
2001).
(x) Epistemological or ontological ‘meta-beliefs’ about how to deter-
mine what one should believe, for example, by affirming empirical,
exegetical, faith-based or intuitionist epistemic rules (Crawford
2002: 19).
Through different configurations of such ideas, ideologies define the key
political differences between individuals and groups, as well as the com-
monalities and unquestioned common sense on which many political
arrangements rest. Critically, political theorists must recognize that
the ideas in question are often generated only unconsciously or
semi-consciously, involve both explicit ‘beliefs’ and more latent
Ideological Analysis 319

‘aliefs’ (Gendler 2008) and might be produced by subtle rhetorical

devices and discursive techniques – the classic focus of critical theor-
ists and discourse analysts. For example, ‘Critical Discourse Analysts’ are
interested in processes like nominalization, the way in which the consistent
rendering of social events as nouns (‘the recession’, ‘urban decay’, ‘home-
lessness’, ‘suppression of the protests’) disguises the human agency
responsible for such events, instead presenting them as natural, agentless,
happenings or default features of the world (Billig 2008). Interrogating
such discursive devices and techniques can be important to analysing how
the ideas that make up ideologies are successfully disseminated, and why
they lead to certain outcomes.
Via such ideas, ideologies critically shape the decision-making and
the patterns of life of whole societies and particular political actors by
shaping what appears desirable and legitimate – whether amongst the
elites who directly make government policies or the ordinary citizens
who may engage in a variety of politically salient activities to contest
or reinforce them. Consequently, they have substantial political out-
comes – producing or affecting the actual and possible political states
of affairs that political theorists analyse, diagnose, critique, alter or
recommend.

5 Conclusion
Ideological analysis is difﬁcult, often tentative and always open to con-
testation and challenge. I have certainly not exhausted key methodologi-
cal precepts and techniques. Nor are the many points of advice or
empirical claims about human thinking that I have asserted immune
from criticism. What this chapter might do, I hope, is two things: to
show political theorists that ideological analysis matters to a far greater
extent than is generally appreciated, and to show them how they can do it
better. An expanded engagement with ideology in political theory is an
exciting direction for the discipline, and suggests a panoply of potential
research projects. But for it to be successful, political theorists need to
draw on interdisciplinary research on ideology more thoroughly, and to
reﬂect on how to study ideology more deeply.

References
Atran, Scott and Axelrod, Robert, 2008. ‘Reframing sacred values’, Negotiation
Journal 24, 221–46.
Ball, Terence, Farr, James and Hanson, Russell L., eds., 1989. Political Innovation
and Conceptual Change. Cambridge University Press.
320 Chapter 14

Baurmann, Michael, 2007. ‘Rational fundamentalism? An explanatory model of

fundamentalist beliefs’, Episteme 4, 150–66.
Bell, Daniel, 1960. The End of Ideology: On the Exhaustion of Political Ideas in the
Fifties. Cambridge, MA: Harvard University Press.
Berger, Peter and Luckmann, Thomas, 1967. The Social Construction of Reality.
New York: Anchor Books.
Berlin, Isaiah, 1969. Four Essays on Liberty. Oxford University Press.
Billig, Michael, 2008. ‘The language of critical discourse analysis: the case of
nominalization’, Discourse and Society 19, 783–800.
Boudon, Raymond, 1989. The Analysis of Ideology. Cambridge: Polity Press.
Boudon, Raymond, 1999. ‘Local vs general ideologies: a normal ingredient of
political life’, Journal of Political Ideologies 4, 141–61.
Carmines, Edward and D’Amico, Nicholas, 2015. ‘The new look in political
ideology research’, Annual Review of Political Science 18, 205–16.
Chandler, David, 2004. ‘The responsibility to protect? Imposing the “liberal
peace”’, International Peacekeeping 11, 59–81.
Constant, Benjamin, 1988. Biancamaria Fontana, ed., Political Writings.
Cambridge University Press.
Coote, Anna, 2014. Who made ideology a dirty word? The NEF Blog [Online].
Available from: www.neweconomics.org/blog/entry/who-made-ideology-a-dirty-
word. Accessed 26 May 2015.
Crawford, Neta, 2002. Argument and Change in World Politics: Ethics,
Decolonization and Humanitarian Intervention. Cambridge University Press.
Doris, John and Murphy, Dominic, 2007. ‘From My Lai to Abu Ghraib: the moral
psychology of atrocity’, Midwest Studies in Philosophy 31, 25–55.
Eagleton, Terry, 1991. Ideology: An Introduction. London: Verso.
Edelman, Murray, 1977. Political Language: Words That Succeed and Policies That
Fail. New York: Academic Press.
Evans, Gareth, 2011. ‘Ethnopolitical conﬂict: when is it right to intervene?’,
Ethnopolitics 10, 115–23.
Evans, Jocelyn, 2002. ‘In defence of Sartori: party system change, voter pre-
ference distribution and other competitive incentives’, Party Politics 8,
155–74.
Fairclough, Norman, 2001. Language and Power. Harlow: Pearson Education.
Fairclough, Norman, 2010. Critical Discourse Analysis: The Critical Study of
Language. Harlow: Pearson Education.
Finlayson, Alan, 2012. ‘Rhetoric and the political theory of ideologies’, Political
Studies 60, 751–67.
Foucault, Michel, 1982. ‘The subject and power’, Critical Inquiry 8, 777–95.
Freeden, Michael, 1996. Ideologies and Political Theory: A Conceptual Approach.
Oxford University Press.
Freeden, Michael, 2005a. ‘Confronting the chimera of a “post-ideological” age’,
Critical Review of International Social and Political Philosophy 8, 247–62.
Freeden, Michael, 2005b. ‘What should the “political” in political theory
explore?’, The Journal of Political Philosophy 13, 113–34.
Ideological Analysis 321

Freeden, Michael, 2006. ‘Ideology and political theory’, Journal of Political

Ideologies 11, 3–22.
Freeden, Michael, ed., 2007. The Meaning of Ideology: Cross-disciplinary
Perspectives. Oxford University Press.
Freeden, Michael, 2008. ‘Thinking politically and thinking about politics:
language, interpretation, and ideology’, in David Leopold and Marc Stears,
eds., Political Theory: Methods and Approaches. Oxford University Press,
196–215.
Freeden, Michael, 2012. ‘Interpretative realism and prescriptive realism’, Journal
of Political Ideologies 17, 1–11.
Freeden, Michael, 2013a. ‘The morphological analysis of ideology’, in
Michael Freeden, Lyman Tower Sargent and Marc Stears, eds., The Oxford
Handbook of Political Ideologies. Oxford University Press, 115–37.
Freeden, Michael, 2013b. The Political Theory of Political Thinking: The Anatomy of
a Practice. Oxford University Press.
Freeden, Michael, Lyman Tower Sargent and Marc Stears, eds., 2013. The Oxford
Handbook of Political Ideologies. Oxford University Press.
Furnham, Adrian, 2003. ‘Belief in a just world: research progress over the past
decade’, Personality and Individual Differences 34, 795–817.
Galston, William, 2010. ‘Realism in political theory’, European Journal of Political
Theory 9, 385–411.
Geertz, Clifford, 1964. ‘Ideology as a cultural system’, in David Apter, ed., Ideology
and Discontent. London: Free Press of Glencoe, 47–76.
Gendler, Tamar Szabó, 2008. ‘Alief and belief’, The Journal of Philosophy 105,
634–63.
George, Alexander and Andrew Bennett, 2005. Case Studies and Theory
Development in the Social Sciences. Cambridge, MA: MIT Press.
Gerring, John, 1997. ‘Ideology: a deﬁnitional analysis’, Political Research Quarterly
50, 957–94.
Geuss, Raymond, 2003. ‘Neither history nor praxis’, European Review 11, 281–92.
Haidt, Jonathan, Jesse Graham and Craig Joseph, 2009. ‘Above and below
left-right: ideological narratives and moral foundations’, Psychological Inquiry 20,
110–19.
Hall, Stuart, 1996. ‘The problem of ideology: Marxism without guarantees’, in
David Morley and Kuan-Hsing Chen, eds., Stuart Hall: Critical Dialogues in
Cultural Studies. London: Routledge, 24–45.
Hamilton, Malcolm, 1987. ‘The elements of the concept of ideology’, Political
Studies 35, 18–38.
Hammack, Philip, 2008. ‘Narrative and the cultural psychology of identity’,
Personality and Social Psychology Review 12, 222–47.
Hardwig, John, 1985. ‘Epistemic dependence’, Journal of Philosophy 82, 335–49.
Hochschild, Jennifer L., 2001. ‘Where you stand depends on what you see: con-
nections among values, perceptions of fact, and political prescriptions’, in
James Kuklinski, ed., Citizens and Politics: Perspectives from Political Psychology.
Cambridge University Press, 313–40.
322 Chapter 14

Howarth, David, Aletta Norval and Yannis Stavrakakis, eds., 2000. Discourse
Theory and Political Analysis: Identities, Hegemonies and Social Change.
Manchester University Press.
Huddy, Leonie, 2001. ‘From social to political identity: a critical examination of
social identity theory’, Political Psychology 22, 127–56.
Humphrey, Matthew, 2005. ‘(De)contesting ideology: the struggle over the mean-
ing of the struggle over meaning’, Critical Review of International Social and
Political Philosophy 8, 225–46.
Ipsos Mori. 2014. Perceptions Are Not Reality: Things the World Gets Wrong.
Available: www.ipsos-mori.com/researchpublications/researcharchive/3466/
Perceptions-are-not-reality-Things-the-world-gets-wrong.aspx. Accessed 9
May 2015.
Jost, John, 2006. ‘The end of the end of ideology’, American Psychologist 61,
651–70.
Jost, John, Christopher Federico and Jaime Napier, 2009a. ‘Political ideology: its
structure, functions and elective affinities’, Annual Review of Psychology 60,
307–37.
Jost, John, Jack Glaser, Arie Kruglanski and Frank Sulloway, 2003. ‘Political
conservatism as motivated social cognition’, Psychological Bulletin 129, 339–75.
Jost, John and Orsolya Hunyady, 2005. ‘Antecedents and consequences of
system-justifying ideologies’, Current Directions in Psychological Science 14,
260–65.
Jost, John, Aaron Kay and Hulda Thorisdottir, 2009b. Social and Psychological
Bases of Ideology and System Justification. Oxford University Press.
Jost, John and Brenda Major, 2001. The Psychology of Legitimacy: Emerging
Perspectives on Ideology, Justice and Intergroup Relations. Cambridge University
Press.
Kahneman, Daniel, 2012. Thinking, Fast and Slow. London: Penguin Books.
Knight, Kathleen, 2006. ‘Transformations in the concept of ideology in the twen-
tieth century’, American Political Science Review 100, 619–26.
Laclau, Ernesto, 2007. ‘Ideology and post-Marxism’, in Michael Freeden, ed.,
The Meaning of Ideology: Cross-disciplinary Perspectives. Oxford University Press,
98–109.
Larrain, Jorge, 1979. The Concept of Ideology. London: Hutchinson and Co.
Leader Maynard, Jonathan, 2013. ‘A map of the field of ideological analysis’,
Journal of Political Ideologies 18, 299–327.
Leader Maynard, Jonathan and Mildenberger, Matto, 2016. ‘Convergence and
divergence in the study of ideology: a critical review’, British Journal of Political
Science. FirstView early online edition.
Mann, Michael 2005. The Dark Side of Democracy: Explaining Ethnic Cleansing.
Cambridge University Press.
Mccauley, Clark and Sophia Moskalenko, 2011. Friction: How Radicalization
Happens to Them and Us. Oxford University Press.
Mclellan, David, 1995. Ideology. Buckingham: Open University Press.
Mill, John Stuart, 2008. On Liberty and Other Essays. Oxford University Press.
Mills, Charles, 1999. The Racial Contract. Ithaca, NY: Cornell University Press.
Ideological Analysis 323

Mullins, Willard, 1972. ‘On the concept of ideology in political science’, American
Political Science Review 66, 498–510.
Nelson, Eric, 2005. ‘Liberty: one concept too many?’, Political Theory 33, 58–78.
Norval, Aletta, 2000. ‘The things we do with words – contemporary approaches to
the analysis of ideology’, British Journal of Political Science 30, 313–46.
Opotow, Susan, 1990. ‘Moral exclusion and injustice: an introduction’, Journal of
Social Issues 46, 1–20.
Pinker, Steven, 1998. How the Mind Works. London: Penguin Books.
Price, Monroe, 2000. ‘Information intervention: Bosnia, the Dayton accords, and
the seizure of broadcasting transmitters’, Cornell International Law Journal 33,
67–112.
Rosen, Michael, 2000. ‘On voluntary servitude and the theory of ideology’,
Constellations 7, 393–407.
Rosenberg, Shawn, 1988. Reason, Ideology and Politics. Cambridge: Polity Press.
Rydgren, Jens, 2009. ‘Beliefs’, in Peter Hadström and Peter Bearman, eds.,
The Oxford Handbook of Analytical Sociology. Oxford University Press, 72–93.
Sagar, Paul, 2014. ‘A broken clock’, The Oxonian Review, 3 June. Available from:
www.oxonianreview.org/wp/a-broken-clock.
Sartori, Giovanni, 1976. Parties and Party Systems: A Framework for Analysis.
Cambridge University Press.
Searle, John, 1995. The Construction of Social Reality. London: Penguin.
Shelby, Tommie, 2003. ‘Ideology, racism, and critical social theory’,
The Philosophical Forum 34, 153–88.
Shenhav, Shaul, 2006. ‘Political narratives and political reality’, International
Political Science Review 27, 245–62.
Simonds, A. P., 1989. ‘Ideological domination and the political information
market’, Theory and Society 18, 181–211.
Skinner, Quentin, 1974. ‘Some problems in the analysis of political thought and
action’, Political Theory 2, 277–303.
Skinner, Quentin, 1998. Liberty Before Liberalism. Cambridge University Press.
Skinner, Quentin, 2002a. ‘A third concept of liberty’, Proceedings of the British
Academy 117, 237–68.
Skinner, Quentin, 2002b. Visions of Politics. Volume I: Regarding Method.
Cambridge University Press.
Steger, Manfred, 2008. The Rise of the Global Imaginary: Political Ideologies from the
French Revolution to the Global War on Terror. Oxford University Press.
Swift, Adam, 2003. How Not to Be a Hypocrite: School Choice for the Morally
Perplexed Parent. London: Routledge.
Tajfel, Henri, 1974. ‘Social identity and intergroup behaviour’, Social Science
Information 13, 65–93.
Thagard, Paul, 2010. ‘Explaining economic crises: are there collective
representations?’, Episteme 7, 266–83.
Thagard, Paul, 2012. ‘Mapping minds across cultures’, in Ron Sun, ed., Grounding
Social Sciences in Cognitive Sciences. Cambridge, MA: MIT Press, 35–62.
Thompson, John, 1984. Studies in the Theory of Ideology. Cambridge: Polity Press.
324 Chapter 14

Tileagă, Christian, 2007. ‘Ideologies of moral exclusion: a critical discursive

reframing of depersonalization, delegitimization and dehumanization’, British
Journal of Social Psychology 46, 717–37.
Tsintsadze-Maass, Eteri and Richard Maass, 2014. ‘Groupthink and terrorist
radicalization’, Terrorism and Political Violence 26, 735–58.
Tully, James, 1983. ‘The pen is a mighty sword: Quentin Skinner’s analysis of
politics’, British Journal of Political Science 13, 489–509.
Tversky, Amos and Daniel Kahneman, 1974. ‘Judgment under uncertainty: heur-
istics and biases’, Science 185, 1124–31.
Valentini, Laura, 2012. ‘Ideal vs. non-ideal theory: a conceptual map’, Philosophy
Compass 7, 654–64.
Van Dijk, Teun, 1998. Ideology: A Multidisciplinary Approach. London: Sage
Publications.
Van Dijk, Teun, 2013. ‘Ideology and discourse’, in Michael Freeden, Lyman
Tower Sargent and Marc Stears, eds., The Oxford Handbook of Political
Ideologies. Oxford University Press, 175–96.
Varki, Ajit and Danny Brower, 2013. Denial: Self-Deception, False Beliefs and the
Origins of the Human Mind. New York: Twelve.
Vigna, Stefano Della and Matthew Gentzkow, 2010. ‘Persuasion: empirical
evidence’, Annual Review of Economics 2, 643–69.
Weber, Max, 2009. ‘Science as a vocation’, in Hans Heinrich Gerth and
C. Wright Mills, eds., From Max Weber: Essays in Sociology. Abingdon:
Routledge, 129–56.
Welsh, Jennifer, 2010. ‘Implementing the “responsibility to protect”: where expec-
tations meet reality’, Ethics and International Affairs 24, 415–30.
Wittgenstein, Ludwig, 2001. Philosophical Investigations. Malden, MA: Blackwell
Publishing.
Wodak, Ruth and Michael Meyer, eds., 2009. Methods of Critical Discourse Analysis.
London: SAGE Publications.
Worsnip, Alex, 2017. ‘Cryptonormative judgements’, European Journal of
Philosophy.
Yanow, Dvora and Peregrine Schwartz-Shea, 2006. Interpretation and Method:
Empirical Research Methods and the Interpretive Turn. Armonk, NY: M.E. Sharpe.
Žižek, Slavoj, ed., 1994a. Mapping Ideology. London: Verso.
Žižek, Slavoj, 1994b. ‘The spectre of ideology’, in Slavoj Žižek, ed., Mapping
Ideology. London: Verso, 1–33.
15 How to Do a Political Theory PhD

Keith Dowding and Robert E. Goodin

The other chapters in this book describe and critique the standard
methods used in analytical political theory. Here we give a few pointers,
of the sort we give to our students, about doing a political theory PhD.

1 Choosing Your Topic

Choose a topic that you actually like. You’re going to have to live with
it for three, four or more years.
Get on with it. Don’t take forever making the choice. You have only
three or four years to write it, and time spent dithering over a topic eats
into writing time.
What’s the problem? A thesis needs to answer a question. It must
address a problem; it is not simply your thoughts on a subject in which
you are interested.
A thesis needs a thesis. You need to settle on more than just a topic or
even a problem: a thesis needs a thesis. It has to argue something.
Why are others wrong? The simplest way of thinking of an answer is to
consider what others have said and why they are wrong.
It’s not your life’s work. Every initial thesis proposal is too big and
ambitious. Remember: you are not Marx, and this is not Kapital. If it is
any good, you will get a job and write other books.

2 Organizing Your Work

Papers or monograph? The traditional thesis is a monograph –
a sustained argument over six to eight chapters. Thesis by papers is now
established in political science and can be transferred to political theory.
But it is not an easy option. A monograph needs only a couple of original
chapters; in a set of papers, each of them needs to say something original.
Bite-sized pieces. Break the thesis down into chapters and focus on one
at a time. Don’t spend too long on each one, but move to the next, even
when you know that more work is required. Once they are all written, you
can ﬁne-tune and work on the ﬁnal draft.

325
326 Chapter 15

Plan, but not too rigidly. You will need a plan for the whole thesis
from the start. But your structure and argument may well change, even
radically, before the end. Don’t worry. Just keep writing, and repackage
later.
Reading is the enemy of writing. At the beginning you will need to
read widely around the literature to get the lie of the land and situate your
work within it. That done, get on with writing; don’t chase up footnotes or
pursue everything written on the subject. Getting your own ideas down
and then seeing who has said something similar is a good way of
distinguishing what you want to say from others. That way you avoid
following the same ruts as the existing literature.
Let it go. Everything can be improved with a little more work, but don’t
let that stand in the way of submission. Plug the big holes and learn to live
with the problems (though plan a response in case someone else, espe-
cially your PhD examiner, spots them).

3 Planning for Life Post-PhD

What is it for? If your aim is to enter the academic profession, writing
a PhD thesis is not enough. You also need to publish, and you should be
thinking about that aspect from the start.
Index

absence of interference, 165 Aristotle, 94–5, 105, 175

abstract reasoning, 5 atheism, 262
abstraction in positive political theory, 196 attachment and ideological analysis,
Ackerly, Brooke, 285 315–17
acontextualism, 313 Austen, Jane, 106
act deontology, 133 authoritarian demagogue, 99
act-utilitarianism, 134 autonomy concept, 114, 181
adaptive reconstruction, 251, 254–5
Adorno, Theodor, 119 background theories, 58, 61
advice in analytical writing, 19–20 Bajpai, Rochana, 283
aesthetic interpretation, 243 balance of power, 143
agonism, 254 Ball, Terence, 249
agreement and realistic idealism, 142–3 Barber, Benjamin, 148
ambiguity of concepts, 156, 255 bargaining theory, 238
anachronistic conceptualizations, 250, Barry, Brian, 86–7
255–6 Begriffsgeschichte (history of concepts),
analytic methodologies, 288 249–50
analytic normative theory, 93 behavioral inference, 312
analytical philosophy, 153 beliefs in ideological analysis, 318
analytical political theory. See also Bell, Daniel, 290
conceptual analysis in political theory; Bentham, Jeremy, 99
political theory Berlin, Isaiah, 119, 155, 250, 254, 307
complementing and competing ideas, Bernasconi, Robert, 258–9
12–13 better chances, defined, 221
conclusion, 13–16 better prizes, defined, 221
defined, 2 bias concerns, 23, 36, 290–1, 310, 316
historical reflections, 4–6 Binmore, Ken, 225
how to write, 18–20 book-length reconstruction, 254–5
introduction, 1–2, 6–7 The British Journal of Political Science, 210
meta-ethics and, 92 British Labour Party (BLP), 117
overview, 8–12 Buchanan, James M., 217
reasons for, 3–4
scope of, 6–8 Cabrera, Luis, 285
Angle, Stephen, 276 Camping Trip thought experiment, 22
Anglo-American political theory, 91–2, 93 cardinal utility functions, 220
anonymous Pareto principle, 76 Carens Market example, 138
anti-Aristotelian philosophical context, 247 Carver, Terrell, 258
anxiety suppression, 316 characterizations in ideological
application approach to models, 210–13 analysis, 318
arbitrary concepts, 171 Christiano, Thomas, 112
Arendt, Hannah, 280 classical liberal perspective, 258
argumentative relevance, 33–4 classification of concepts, 156–7

327
328 Index

cluster concepts, 165 Confucian Democracy in East Asia

co-authorship, 291 (Sungmoon Kim), 289
Coetzee, J.M., 106 Confucianism, 276, 277
Cognitive-Affective Mapping connotations of concepts, 157
methodology, 315 conscience, 5
cognitive dissonance minimization, 316 consent contractualism, 65, 67, 68–9, 70–3
cognitive efficiency concerns, 316 conservatism and realistic idealism, 141
Cohen, G.A., 22, 125, 137, 140, 158–9 Considerations on the Government of Poland
coherentism, 50 (Rousseau), 250
collective ideologies, 302 consistency bias, 310, 316
Commission on Social Justice (BLP), 117 Constant, Benjamin, 310
common knowledge assumption, 227 constituency of fairness contractualism, 74
comparative political thought (CPT) The Constitution of Equality: Democratic
conclusions, 291–2 Authority and Its Limits
critical approach to, 285–6 (Christiano), 112
defined, 273–5 Contemporary Political Philosophy
historical approach to, 278–82 (Kymlicka), 112
interpretive approach to, 282–4 content in ideological analysis, 312–15
introduction, 4, 11, 270–3 context-specific questions for political
methodology, 275–6, 286–91 theory, 308
normative-analytic approach to, contextualism, 244, 245–8, 261, 262, 306,
276–8 313–14
complex notion, 28, 163 continental political theory, 6
compound lotteries, 221 contingency of thought experiments, 28–30
comprehensibility motives, 316 contingently hypothetical thought
compulsory voting systems, 210, 211 experiments, 38–9
conception of justice, 136 continuity, defined, 220–1
conceptual analysis in political theory contractarianism, 234, 235, 238
concepts of, 154–63 contractual scenario, 65
conceptual change, 161–3 contractualism
conceptual connections, 167–8 consent contractualism, 65, 67,
conclusion, 186–7 68–9, 70–3
design of, 170–2 designing of, 70
as disambiguation, 168–70 introduction, 6, 8–9, 65–6
essentially contested concepts, 176–7 misuse of, 82–8
essentially/non-essentially evaluative moral/political problems and, 83–4, 87, 89
concepts, 172–6 normative content in, 86–8
as extensional analysis, 164–7, 174, rationality contractualism, 79–82
183, 185 reasons for, 66–70
introduction, 10, 12, 153–4 types of, 85–6, 88
normative implications, 179–82 use of, 88–9
overview, 163–72, 183–6 wrongness and injustice, 82–3, 89
political concepts, 172–9 cooperative game theory, 230
principles and, 159–61 core normative theory, 218
as resolution, 163–4 correlated equilibrium, 228
terms and, 154–9 corruption and realistic idealism, 148–9
value-neutrality and ideology, 177–9 Cosmopolitain War (Fabre), 112
conceptual change, 161–3 creative political theory, 310–11
conceptual connections, 167–8 creative writing, 105
conceptual objection against criteria of application, 160, 166
utilitarianism, 168 critical approach to CPT, 285–6
conceptualization, 264 Critical Discourse Analysts, 319
conflict and social theory, 151 cruel and unusual punishment, 166
conflict management, 145–8 cryptonormativity, 300
conflicts of interest and equilibrium, 56, 59 cultural differences, 275
Index 329

Dallmayr, Fred, 273, 280 exemplary realism, 113–17

data sources, 282 expected behavior, 318
De Cive (Hobbes), 253–4 expected utility, 222
decision-making and equilibrium, 56 expected utility calculus, 223–4
decision theory, 236–9 expected utility property, 224
democratic decision procedures, 177 Experience Machine thought experiment,
deontology, 6, 193 22–3, 29, 36
descriptive hypothetical examples, 24, 25–6 explanatory principles, 38–9
desirability and realistic idealism, 138–9 extension concept, 157
development of methodology, 288 extensional analysis, as conceptual analysis,
devices of representation, 58 164–7, 174, 183, 185
didactic discourse, 97 extreme narrative-framing constraint, 36
differentiating definitions, 169
dimuqratiyyat al-khubs (democracy of Fabre, Cecile, 112
bread), 283 fairness contractualism, 3, 65, 68–9,
disambiguation, as conceptual analysis, 82–3, 89
168–70 false consciousness, 301
discernible defect, 139–40 family resemblances in games, 165
discursive saturation, 316 Famous Violinist thought experiment, 43
disputes over concepts, 158 Farr, James, 249
distinction importance, 13, 160–1 Fat Man on the Bridge thought
distinctions in analytical writing, 18 experiment, 43
distinctive idiosyncratic worldviews, 298 Faust (Goethe), 117
distributive justice, 146, 160, 162, 164, 170 feasibility and realistic idealism, 138–9
diversity concept, 132, 141–3, 181, 305 fee-paying schools, 305
dominant strategy, 230 feminism, 264
Douglass, Robin, 259 Feminist Philosophers blog, 7
Downsian model, 204, 205f, 205, 208, 209 fiction writing, 105–6
Dunn, John, 113, 256, 257 field of possibilities, 318
first-order theorizing, 113
Eagleton, Terry, 301 fleshed-out alternatives constraint, 35
East Meets West (Bell), 290 Foot, Philippa, 21–2
économistes, 259 formal political theory, 194
egalitarianism, 61, 124 Foucault, Michel, 252–3, 259
Elster, Jon, 258 foundation follows function, 146
emotional attachment, 318 foundationalism, 50, 55
emotional engagements, 91 free association in analytical writing, 19
empirical political science, 47 free-market ideologies, 304
empirical reconstruction, 251, 255 Freeden, Michael, 272, 290
empirical research, 14, 15, 297 freedom
Enemy in the Mirror (Euben), 279–80, 287 determination of amount, 182
epistemic dependence, 316 money and, 158–9
epistemic motives, 315 non-essentially evaluative concepts,
epistemic value, 49 172–3
epistemological meta-beliefs, 318 of religion, 142
equal bargaining power, 66 self-realization and, 177
esotericism, 259, 261 of speech, 142, 307
essentially contested concepts, 176–7
essentially evaluative concepts, 172–6 Galston, William, 113, 181
ethical-political thought, 97, 116, 270 game-theoretic political theory, 194–5, 238
Euben, Roxanne, 273, 279–80, 283, Gauthier, David, 217, 237
287, 288 genealogy of concepts, 182
Eurocentrism, 273 genealogy of models, 203
evaluative concepts, 172–6 Gerring, John, 300
exceptionalism, 300 Geschichtliche Grundbegriffe, 249
330 Index

Geuss, Raymond, 119, 308 integrative approach, 311–12

Gilabert, Pablo, 140 introduction, 11–12, 297–8
Gintis, Herbert, 233–6 meaning of, 298–303
globalization, 273 normative political theory, 298, 311
Godrej, Farah, 280 political theory and, 305–9
Goldilocks solution, 76 structure of ideas, 315
good practice, core principles, 261–4 ideological incentivization, 316
Goodman, Nelson, 46 ideologizing, defined, 302
Gray, John, 251, 254 imaginability of thought experiments,
Green, Donald, 232 28–30
Grotius, Hugo, 162–3 imaginative opacity, 39–40
grounded political theory, 285 impartiality model, 239
groupthink, 316 impassioned practice
empathetic consideration of particulars,
Hamilton, Malcom, 300 102–3
Hamlin, Alan, 135 moral sentimentalism and, 93–6
Hansen, Russell, 249 objections to, 96–100
Harsanyi, John, 217, 229, 239 perspectives on, 103–4
high liberalism, 121 psychologically realistic fiction, 104–6
Hinman v. Pacific Air Transport (1936), show, don’t tell, 101–2
146–7 simple, ordinary language usage, 106–8
historical approach to CPT, 278–82 techniques for, 100–8
Hobbes, Thomas writing interesting stories, 100–1
account of representation, 257 impersonality model, 239
ambiguous references, 253–4 important facts in analytical writing, 18
atheism, 262 In the Beginning Was the Deed (Williams),
Leviathan, 183 113, 117
natural right and, 183–4 incoherent concepts, 158, 171
political theory and, 236 independent credibility, 48
republican context, 259 inferential justifying of beliefs, 50
holistic self-scrutiny, 93 informal positive political theory, 195
Hollis, Martin, 258 information in fairness contractualism, 76
Homo economicus concept, 157 information in rationality
Honig, Bonnie, 113 contractualism, 81
human states of mind, 106 injustice, 82–3, 89, 132
humanitarian intervention, 306, 308 inquiry methods, 312
Hume, David, 91, 96–7, 99, 105, 108 integrative approach to ideological analysis,
humility importance, 291 311–12
Humphrey, Matthew, 299 intellectual contexts of texts, 247
hybrid models of contractualism, 86 intelligibility and rational choice
theory, 233
idealization forms, 81 interests in fairness contractualism,
idealization in positive political theory, 196 71, 75–6
idealized problems, 136 interests in rationality contractualism, 80
ideational explanation, 283–4 International Commission on Intervention
ideational resources, 316 and State Sovereignty (ICISS), 306
ideological analysis international political theory, 7
analyzing effects of, 317–19 interpretative analysis, 185
assumptions behind arguments, 309–10 interpreting texts
attachment and, 315–17 Begriffsgeschichte, 249–50
conclusion, 319 conclusion, 264–5
creative political theory, 310–11 contextualism, 244, 245–8
evaluation of institutions, 304–5 core principles of good practice, 261–4
finding content, 312–15 introduction, 11, 12–13, 243–5
idealization forms, 81 reading between/outside the lines, 260–1
Index 331

reconstruction, 251–7 Lesser, Harry, 264

theoretical and normative perspectives, Lever, Annabelle, 210, 211
258–60 Leviathan (Hobbes), 183
interpretive approach to CPT, 282–4 liberalism
interpretive clarity, 186 classical liberal perspective, 258
intra-disciplinary methodological high liberalism, 121
disagreements, 287 legitimacy and realism, 118
introspective psychological experiments, 26 modernity and, 120–1
intuition pumps in analytical writing, 19 reconstructions, 254
intuitionism, 5 libertarian perspective, 258
involuntary disadvantage, 62 linear justifying of beliefs, 50
Islamic extremism, 273, 280 linguistic contexts of texts, 247
Locke, John, 119, 180, 183, 184, 251,
James, Henry, 106 255, 256
Jenco, Leigh, 280 logic of inference, 1
Jim and the Indians thought experiment, 43 Losonsky, Michael, 264
judges and test of theory, 147–8 luck egalitarianism, 61, 62
judgments, 27, 46, 47–9
just-world thinking, 307 Machiavelli, Niccolò, 252, 260, 262
justice MacIntyre, Alastair, 280
conceptual structure of, 175–6 Macpherson, C.B., 258
defined, 163, 175 March, Andrew, 277–8, 287
ordinary justice, 170 Marmot, Michael, 125
principles of justice, 136–8 Marx, Karl, 258
realistic idealism and, 144 Marxism, 244, 258, 259, 305
standards of justice, 175 McCormick, John, 258
justified slavery concepts, 157 Median Voter Theorem, 201
mediators and test of theory, 147–8
Kant, Immanuel, 180–1 Melzer, Arthur, 259, 261
Keating, Christine, 285 mental visualization, 25
key devices in analytical writing, 18–19 meta-beliefs, 318
key objectives, 318 meta-ethics and analytical political
KISS (‘Keep It Simple, Stupid’), 18 theory, 92
Kleimt, Hartmut, 233 metaphorical abstraction, 302
Kundera, Milan, 106 methodological publications, 16
Kymlicka, Will, 112 methodological reflection, 15
methodology
laissez-faire principle, 179 analytic methodologies, 288
language skills, 289 Cognitive-Affective Mapping
language usage in impassioned practice, methodology, 315
106–8 comparative political thought, 275–6,
Laplacean principle of indifference, 225 286–91
Latin grammatical structures, 107 development of, 288
law, defined, 174 intra-disciplinary methodological
layers of analytical writing, 18 disagreements, 287
Lectures on Rhetoric and Belles Lettres Oppenheim’s methodological
(Smith), 106 project, 173
Left/Right political concepts, 156 pluralism in, 270, 286
legal positivism, 167 qualitative methodology, 290
legitimacy range of, 272
demand, 115 The Methods of Ethics (Sidgwick), 252
oppression-legitimating systems, 301 Mill, James, 250
political legitimacy, 123 Mill, John Stuart, 119, 153, 175, 243,
realism and, 118–25 263, 307
realistic egalitarianism, 118–25 mimicking situations, 232
332 Index

models non-essentially evaluative concepts, 172–6

application approach, 210–13 non-ideal political theory, 307, 308
basic relationships of, 204–5 non-political texts, 246
conclusion, 213 non-question-beggingness, 30–1
genealogy of, 203 non-verbal expressions, 279
limitations of, 209 normative-analytic approach to CPT,
motivational assumptions, 207, 208 276–8
purpose of, 205–7 normative argument, 168, 288
usefulness vs. realism, 196–7 normative case studies, 98, 102
variations on, 208 normative concepts, 159
zooming in/out, 209–10 normative content in contractualism, 86–8
moderate narrative-framing normative contractarian project, 237
constraint, 35–7 normative implications of conceptual
modernity analysis, 179–82
interpretations of, 121–2 normative perspectives on interpreting
liberalism and, 120–1 texts, 258–60
monarchy concept, 155 normative political theory
moral problems and contractualism, 83–4, ideological analysis, 298, 311
87, 89 introduction, 7, 47
moral requirements, 27 prohibitions/obligations, 318
moral sentimentalism rational choice theory as, 233–6
impassioned practice and, 93–6 sentimentalism and, 91–3
introduction, 9, 12, 91–3 Norval, Aletta, 300
objection against manipulation, 99–100 notion of reasonableness, 76
objections to impassioned practice, Nozick, Robert, 22–3, 29, 43–4, 112
96–100
techniques for impassioned practice, objection against manipulation, 99–100
100–8 Objection of Bias, 23
moral theory, 54, 56, 134–5, 251, 255 Objection of Inherent Ambiguity, 23, 24
moralized definitions of political Okin, Susan, 258
concepts, 174 On Liberty (Mill), 243, 253, 262
Morals by Agreement (Gauthier), 217 ontological meta-beliefs, 318
Morganthau, Hans, 183 ontological perspective, 287
motivations, 194, 197–8, 207, 208 Oppenheim’s methodological project, 173
motives in fairness contractualism, 71, 75–6 oppression-legitimating systems, 301
motives in rationality contractualism, 80 ordinal utility function, 219
Mouffe, Chantal, 113 ordinary justice, 170
multiparty election systems, 304 organizing analytical writing, 18
multiple equilibria, 229 Original Position thought experiment,
multivalent words, 168–9 22, 29
outcome egalitarianism, 61
narrative bias, 36 outcomes and patterns of cooperation, 134
narrative discourse, 97 Oxford English Dictionary, 169, 184
Nash equilibrium, 228, 229, 230
navigation easements, 147 parametric choice, 222–6, 226f, 226t
Nazi concentration camp experiments, 2 passionate reasoning, 264
necessity of thought experiments, 28–30 Pathologies of Rational Choice Theory
Nelson, Matthew, 283–4 (Shapiro, Green), 232
neuroscientific methods, 312 patterns of cooperation, 134
nominalization process, 319 pejorative connotations, 301
nomothetic principles, 97–8 personal distress and equilibrium, 56
non-agreement in contractualism, 73, personal ideologies, 302
77–8, 81–2 persuasive definitions, 182
non-cooperative game theory, 230 Philip, Mark, 113
non-domination idea, 162, 167 Phillips, Anne, 182
Index 333

philosophical evidence, 262–3 positivism, 193

philosophical interpretation, 243, 255 possessive individualism, 258
philosophical respectability, 30–3 preference maximization theory, 218–22,
Philosophy and Real Politics (Adorno), 119 219t, 222t
Pickett, Kate, 125 preferences, defined, 220
placement in analytical writing, 20 primary question in fairness contractualism,
Plato, 4, 264 71, 74–5
Poetics (Aristotle), 94–5 The Prince (Machiavelli), 252, 262
political authority claims, 114 principles, defined, 46
political concepts, conceptual analysis, principles and concepts, 159–61
172–9 principles of justice, 136–8
political ideology, 300, 304 prioritarianism, 62
Political Innovation and Conceptual Change priority rules, 59
(Ball, Farr, Hanson), 249 problem-driven research, 14
political legitimacy, 123 proper behavior, 318
political morality, 115–16, 123–4 proper members of society/group, 317
political philosophy, 179–80 property of rights, 256
Political Philosophy: A Beginner’s Guide for property rights, 146, 180, 184
Students and Politicians (Swift), 112 proto-thought experiments, 25
political problems and contractualism, prototypical behavior, 318
83–4, 89 provisional definition, 185
political theory. See also analytical political psychological experiments, 24, 26–8
theory; comparative political thought; psychological motives, 315
conceptual analysis in political theory; psychologically realistic fiction, 104–6
positive political theory
assumptions behind arguments, 309–10 qualitative methodology, 290
context-specific questions for, 308 quantitative empirical work, 212
continental political theory, 6 quantitative research, 206
creative political theory, 310–11 question-driven political theory, 286
doctorate study, 325–6 Qutb, Sayyid, 283, 287, 288
forms of discourse, 108
grounded political theory, 285 race, logical vs. theoretical concept, 179
ideological analysis and, 305–9 racial profiling, 305
international political theory, 7 racial subordination, 52
introduction, 12 radical uncertainty, 225
non-ideal political theory, 307 random killing judgment, 53
normative considerations, 178 rational choice theory
question-driven political theory, 286 decision theory, 236–9
real political theory, 96 intelligibility and, 233
realism, 3, 119, 124–5, 126–7 introduction, 10–11, 217–18
realism and, 124–5, 126–7 as normative theory, 233–6
realistic idealism, 149 parametric choice, 222–6, 226f, 226t
politics and nomothetic principles, 97 political theory and, 194
politics and subjective experiences, 104 preferences and utility functions, 218–22,
politics of decision-making, 132–41 219t, 222t
politiques, 259 strategic choice, 226–31, 227f, 227t,
Pond and Envelope thought 228f, 228t, 229f, 229t, 231f, 231t
experiment, 32–3 unrationality of people, 231–3
positive liberty, 307 rationalism, 287
positive political theory rationality contractualism, 3, 65, 69–70,
defined, 192–9 79–82
extended example of, 199–203, 200f Rawls, John
guide to, 203–10, 204t, 205f conception of justice, 136, 137, 162
introduction, 10, 12, 192 criteria of application, 160
models, 212 distinction importance, 160–1
334 Index

Rawls, John (cont.) religious intolerance, 52

distributive justice, 160 Republic (Plato), 4, 263
fairness contractualism, 85 research question, 286–7
interpretation of modernity, 121–2 resolution, as conceptual analysis, 163–4
justice as a peak, 144 resources and freedom, 155
overview, 15, 22, 29, 46, 50 responsibility-catering prioritarianism, 62
rational choice theory, 217–18, 239 revising analytical writing, 18
realistic idealism, 135, 148 Rhetoric (Aristotle), 94–5
social contract theory, 235 rhetorical discourse, 97
social cooperation for mutual rhetorical presentation, 316
advantage, 85 Rich and Superrich thought
A Theory of Justice, 46, 50, 57 experiment, 40–1
Raz, Joseph, 162 risk neutrality, 224
reading between/outside the lines, 260–1 rival conceptions, 161
real philosophy, 96 Rosenblatt, Helena, 247–8
real political theory, 96 Rousseau, Jean-Jacques, 5, 246, 247–8, 259
real-world ideologies, 306 rudimentary alternative constraint, 34–5
realism rule-consequentialist interpretation, 255
exemplary realism, 113–17 Russell, Bertrand, 103
introduction, 9, 12, 112–13
legitimacy and realistic egalitarianism, Sapiro, Virginia, 264
118–25 Schelling, Thomas, 230
political theory, 3, 119, 124–5, 126–7 Scottish Enlightenment, 134
summary, 126–8 self-esteem motives, 316
realism and moralism in political theory self-ownership, 169, 180–1
(RMPT), 113, 126 self-realization, 169, 177
realistic egalitarianism, 118–25 Selton, Reinhard, 229
realistic idealism, 4 semantic holism, 40
agreement and, 142–3 semi-autonomous traditions, 281
conclusion, 150–1 sentimentalism. See moral sentimentalism
conflict management, 145–8 sentimentalism in political theory, 3
conservatism and, 141 Shapiro, Ian, 232
corruption, 148–9 shari’a law, 283
diversity and, 132, 141–3 show, don’t tell as impassioned practice,
feasibility and desirability, 138–9 101–2
introduction, 10, 12, 131–2 Simmons, A.J., 251, 254–5, 256
justice and, 144 simple thought experiments, 28
peaks vs. pits, 144–5 simplification in positive political
politics of, 132–41 theory, 196
principles of justice, 136–8 Singer, Peter, 133–4
standards and, 139–40 Skinner, Quentin
strategy of, 135–6 acontextualism, 313
reciprocity principle, 134 ideological perspective, 310
reconstruction, 251–7 introduction, 1, 15
reduction of compound lotteries, 221 republican perspective of, 258, 259
referendum studies, 199–203, 200f, 204t substantive interpretation, 245–6
reflective equilibrium textual interpretation, 243
challenges, 52–6 Smith, Adam, 91
example of, 59–62 The Social Contract (Rousseau), 5, 246
how to use, 56–9 social contract theory, 235
introduction, 3, 8, 12, 46–7 social epidemiology, 125
judgments, 46, 47–9 social institutions, 144
wide reflective equilibrium, 49–52, 51t social justice, 164
refugee persecution example, 170–1 social processes, 315
religio-political identities, 283 social theory, 132, 151
Index 335

solipsism in theory, 133–5, 136–8 moderate narrative-framing

solitary confinement study, 103 constraint, 35–7
Sorensen, Roy, 35 necessary hypotheticals, 40–1
standards and realistic idealism, 139–40 non-question-beggingness, 30–1
standards of justice, 175 philosophical respectability, 30–3
STEM fields, 107 psychological experiments, 24, 26–8
Stevenson, C.L., 182 rudimentary alternative constraint, 34–5
stipulative error, 299 simple vs. complex, 28
strategic choice, 223, 227f, 227t, 228f, 228t, validity of, 31–3
229f, 229t, 231f, 231t, 226–31 wacky thought experiments, 37–8
strategic deontology, 133 well-posed thought experiments, 30
Strauss, Leo, 261 Ticking Bomb thought experiment, 23,
Straussianism, 244, 261 26, 37
strict preferences, defined, 220 Timmons, Jeffrey, 212
struggles of politics, 285–6 transitive preferences, defined, 220
Suárez, Francisco, 167, 168 triangulation, 291
substantive interpretations, 245 Tripp, Charles, 282–3
substantive objection against Trolley Problem and Transplant Case, 22,
utilitarianism, 168 28, 34, 36, 83–4
substantive research, 16 Trolly thought experiment, 21, 28
Sungmoon Kim, 289 true primitive preferences, defined, 220
Swift, Adam, 112 Truth and Truthfulness (Williams), 118
system justification, 307 Tuck, Richard, 248
systematic analysis, 185 two-party election systems, 304
systematic marginalization, 170 Two Treatises (Locke), 183
systematic reconstruction, 251 tyranny concept, 155
systems of ideas, 298–9
uncertainty principle, 224, 262
Taylor, Charles, 280 under-determination principle, 261
teleportation in thought experiments, 36 unfair equality concept, 41
terminological consistency, 185–6 universal human rights, 271
textual analysis, 312 universalism of political morality, 115,
textual evidence, 262 116–17, 118
Thagard, Paul, 315 unrationality of people, 231–3
theoretical consistency, 186 Unwilling Rock thought experiments, 35
theoretical perspectives on interpreting usefulness idea, 198–9
texts, 258–60 utilitarianism, 6, 50, 82–3, 114, 134, 168
Theory of Games and Economic Behavior (von Utilitarianism (Gaus), 175
Neumann, Morgenstern), 217 utility maximization, 4
A Theory of Justice (Rawls), 46, 50, 57, 235 Utility Monster thought experiment, 23, 29
The Theory of Moral Sentiments (Smith), 106 utility theory, 219t, 222t, 218–22
Thomson, Judith Jarvis, 21–2, 43 utopianism, 119
‘Thou shalt nots’ in justice, 145
thought experiments vague concepts, 157–8
argumentative relevance, 33–4 validity of thought experiments, 31–3
conclusion, 42 value-free definition, 173
contingently hypothetical thought value-neutral definition, 173
experiments, 38–9 value-neutrality and ideology, 177–9
defined, 24–5 value-pluralism, 254
descriptive hypothetical examples vs., Verstehen understanding, 98
24, 25–6 virtú, defined, 252, 260
examples of, 43–4 voluntary voting systems, 210
final distinctions, 28–30 von Neumann, John, 217, 227
imaginative opacity, 39–40 voting behavior study, 197, 199–203,
introduction, 2, 8, 12, 21–4 200f, 211
336 Index

wacky thought experiments, 37–8 Wilt Chamberlain thought

Wallace, David Foster, 106 experiment, 43–4
Weimar Republic, 121 Wolff, Robert Paul, 148–9
well-posed thought experiments, 30 Women in Philosophy (Hutchison, Jenkins), 8
Welsh, Jennifer, 306 Wootton, David, 264
Westernization, 287 writing interesting stories, 100–1
wide reﬂective equilibrium, 49–52, 51t wrongness and contractualism, 82–3, 89
Wilkinson, Richard, 125
Williams, Bernard, 43, 112–17, 126 zero-sum games, 226, 227

Held David Modelos de Democracia
No ratings yet
Held David Modelos de Democracia
154 pages
Great Debates in Jurisprudence Great Debates in Law 9 2nd Edition Nicholas J Mcbride Sandy Steel
No ratings yet
Great Debates in Jurisprudence Great Debates in Law 9 2nd Edition Nicholas J Mcbride Sandy Steel
68 pages
O&#39 Connor, James. La Crisis Fiscal Del Estado. Ediciones Península. 1994. Introducción.
100% (1)
O&#39 Connor, James. La Crisis Fiscal Del Estado. Ediciones Península. 1994. Introducción.
44 pages
LEVITSKY, Steven & Lucan Way, Competitive
No ratings yet
LEVITSKY, Steven & Lucan Way, Competitive
20 pages
Dissertation Topics On Public International Law
100% (2)
Dissertation Topics On Public International Law
8 pages
Richard Kroner - Kant's Weltanschauung PDF
100% (1)
Richard Kroner - Kant's Weltanschauung PDF
140 pages
Campbell, Converse, MIller & Stokes - The American Voter - 1960
No ratings yet
Campbell, Converse, MIller & Stokes - The American Voter - 1960
12 pages
By The Book
No ratings yet
By The Book
2 pages
HENRY - How Modern Democracies Are Shaping Evaluation and The Emerging Challenges For Evaluation
No ratings yet
HENRY - How Modern Democracies Are Shaping Evaluation and The Emerging Challenges For Evaluation
11 pages
UNWTO - TedQual Self Evaluation A1
No ratings yet
UNWTO - TedQual Self Evaluation A1
15 pages
Ethics Case Study Assignment-Erica Kovarik
100% (1)
Ethics Case Study Assignment-Erica Kovarik
8 pages
Assignment On Introduction To Professional Social Work
100% (1)
Assignment On Introduction To Professional Social Work
8 pages
Sandra Halperin and Oliver Heath - Political Research - Methods and Practical Skills (2012)
No ratings yet
Sandra Halperin and Oliver Heath - Political Research - Methods and Practical Skills (2012)
460 pages
Public Administration Theory
No ratings yet
Public Administration Theory
52 pages
AUSIT Code of Ethics and Code of Conduct
0% (1)
AUSIT Code of Ethics and Code of Conduct
4 pages
(GEETHIC) Learning Output 3 - Scribd
No ratings yet
(GEETHIC) Learning Output 3 - Scribd
2 pages
Diss W3 - M2: Branches of Humanities Nature and Functions
100% (1)
Diss W3 - M2: Branches of Humanities Nature and Functions
3 pages
Gabriel Noje The Sacrement of Serving Our Fellow Men
No ratings yet
Gabriel Noje The Sacrement of Serving Our Fellow Men
6 pages
Teune MacroTheoreticalApproaches 1977
No ratings yet
Teune MacroTheoreticalApproaches 1977
13 pages
Module 7
No ratings yet
Module 7
6 pages
The Protestant Work Ethic (Weber)
67% (3)
The Protestant Work Ethic (Weber)
349 pages
Lesson 1 - Philosophy
No ratings yet
Lesson 1 - Philosophy
13 pages
Social Responsibilities & Ethics in Strategic Management
No ratings yet
Social Responsibilities & Ethics in Strategic Management
10 pages
Title of Project Report
No ratings yet
Title of Project Report
18 pages
Prelim Tq-Govsor
No ratings yet
Prelim Tq-Govsor
3 pages
You'Ll Be The Death of Me
No ratings yet
You'Ll Be The Death of Me
348 pages
Derecho y Justicia Ralf Dreier
100% (1)
Derecho y Justicia Ralf Dreier
24 pages
Political Science Document
No ratings yet
Political Science Document
23 pages
JSTOR - Social Psychology and Economics in Environmental Research - C L Spash & A Biel (, 2002)
No ratings yet
JSTOR - Social Psychology and Economics in Environmental Research - C L Spash & A Biel (, 2002)
5 pages
Diminished Parties: Democratic Representation in Contemporary Latin America
No ratings yet
Diminished Parties: Democratic Representation in Contemporary Latin America
369 pages
Pub - Values Ethics and Health Care
No ratings yet
Pub - Values Ethics and Health Care
169 pages
Political Science
No ratings yet
Political Science
40 pages
Leadership Ethics
No ratings yet
Leadership Ethics
25 pages
Ethics Cross Walk 210407
No ratings yet
Ethics Cross Walk 210407
4 pages
Fischer2015 PDF
100% (1)
Fischer2015 PDF
535 pages
Cep Lesson 2weebly
No ratings yet
Cep Lesson 2weebly
5 pages
Comportamiento Politico y Electoral
100% (1)
Comportamiento Politico y Electoral
21 pages
Between X and Y How Process Tracing Contributes To Opening The Black Box of Causality
100% (1)
Between X and Y How Process Tracing Contributes To Opening The Black Box of Causality
19 pages
Thomas J. Sullivan - Introduction To Social Problems-Pearson (2015)
No ratings yet
Thomas J. Sullivan - Introduction To Social Problems-Pearson (2015)
501 pages
Ethics of Abortion PDF
No ratings yet
Ethics of Abortion PDF
2 pages
Distinctive Challenges of Library-Based Interdisciplinary Research and Writing: A Guide
No ratings yet
Distinctive Challenges of Library-Based Interdisciplinary Research and Writing: A Guide
27 pages
Behaviorismo en La Ciencia Política
No ratings yet
Behaviorismo en La Ciencia Política
37 pages
Callen Institutional Corruption and Election Fraud Evidence From A Field Experiment in Afghanistan
No ratings yet
Callen Institutional Corruption and Election Fraud Evidence From A Field Experiment in Afghanistan
41 pages
Gerring - Metodologia de Las Ciencias Sociales - Cap1 PDF
No ratings yet
Gerring - Metodologia de Las Ciencias Sociales - Cap1 PDF
12 pages
The Eroticism of in Equality 0
No ratings yet
The Eroticism of in Equality 0
9 pages
Annotated Bibliography Bailey
No ratings yet
Annotated Bibliography Bailey
6 pages
ZBW 1528327160
100% (1)
ZBW 1528327160
6 pages
ANDUIZA - 2012 - Digital Media and Political Engagement
No ratings yet
ANDUIZA - 2012 - Digital Media and Political Engagement
245 pages
Oxford Handbooks Online: Quantitative Methodology
No ratings yet
Oxford Handbooks Online: Quantitative Methodology
21 pages
U3. Amadeo - Framing Modelo para Armar
No ratings yet
U3. Amadeo - Framing Modelo para Armar
29 pages
Lührmann & Rooney April 2020 PDF
No ratings yet
Lührmann & Rooney April 2020 PDF
38 pages
Qualitative Methods and The Study of Civil War Course Syllabus
100% (1)
Qualitative Methods and The Study of Civil War Course Syllabus
6 pages
Sweeney - Democratic Backsliding, Polarization, Poland
No ratings yet
Sweeney - Democratic Backsliding, Polarization, Poland
47 pages
Easton The Political System (1953)
No ratings yet
Easton The Political System (1953)
352 pages
Sovacool 2016 - How Long Will It Take
No ratings yet
Sovacool 2016 - How Long Will It Take
14 pages
Richard S. Katz Bernhard Wessels The European PBook4You
No ratings yet
Richard S. Katz Bernhard Wessels The European PBook4You
295 pages
Bruce Ackerman PDF
No ratings yet
Bruce Ackerman PDF
10 pages
IgnaziPP1996549 66
No ratings yet
IgnaziPP1996549 66
19 pages
3 Cepiku 2020 04 01 15 39 58
No ratings yet
3 Cepiku 2020 04 01 15 39 58
17 pages
Esteban Ruizponce Madrid - Defining Institutions As Rules
No ratings yet
Esteban Ruizponce Madrid - Defining Institutions As Rules
27 pages
Postmodernism, Civic Engagement
No ratings yet
Postmodernism, Civic Engagement
21 pages
Mahoney 2008 Methods Syllabus
No ratings yet
Mahoney 2008 Methods Syllabus
6 pages
Political Campaigns Changes Voters' Perspective: Abuyog Community College Senior High School Abuyog, Leyte
No ratings yet
Political Campaigns Changes Voters' Perspective: Abuyog Community College Senior High School Abuyog, Leyte
5 pages
ABED Module 3 - Moral Minefield
No ratings yet
ABED Module 3 - Moral Minefield
2 pages
Amartya Sen Approach in Human Development
No ratings yet
Amartya Sen Approach in Human Development
17 pages
Hajer Authoritative Governance Policy Making in The Age of Mediatization
No ratings yet
Hajer Authoritative Governance Policy Making in The Age of Mediatization
219 pages
Pierson - Increasing Returns, Path Dependence, and Study of Politics
No ratings yet
Pierson - Increasing Returns, Path Dependence, and Study of Politics
18 pages
Sartori Comparing and Miscomparing Kopi
100% (1)
Sartori Comparing and Miscomparing Kopi
15 pages
Seeking An Understanding of Poverty That Recognizes Rural-Urban Differences and Rural-Urban Linkages
No ratings yet
Seeking An Understanding of Poverty That Recognizes Rural-Urban Differences and Rural-Urban Linkages
19 pages
Public Policy-Compressed
No ratings yet
Public Policy-Compressed
25 pages
Ikenberry - After Victory (Ch. 1 2 3)
No ratings yet
Ikenberry - After Victory (Ch. 1 2 3)
10 pages
State Capture Western Balkans 2017
No ratings yet
State Capture Western Balkans 2017
69 pages
Dearborn Politics and Public Policy Syllabus Xpxisp
100% (1)
Dearborn Politics and Public Policy Syllabus Xpxisp
5 pages
Pierson 1993 When Effect Becomes Cause
No ratings yet
Pierson 1993 When Effect Becomes Cause
35 pages
Gingrich Häusermann 2015
No ratings yet
Gingrich Häusermann 2015
26 pages
Jasonoff TechHumility
No ratings yet
Jasonoff TechHumility
1 page
Diarmuid Maguire Marxism 2010
100% (1)
Diarmuid Maguire Marxism 2010
10 pages
Paudel (2009) A Critical Account of Implementation Theories
No ratings yet
Paudel (2009) A Critical Account of Implementation Theories
19 pages
Chapter 2 PDF
No ratings yet
Chapter 2 PDF
35 pages
Vivien Schmidt 2012 Discursive Institutionalism
No ratings yet
Vivien Schmidt 2012 Discursive Institutionalism
25 pages
Teaching and Learning Ontology and Epistemology in Political Science.
No ratings yet
Teaching and Learning Ontology and Epistemology in Political Science.
9 pages
Beyond Beyond Policy Analysis: Emergent Policy and The Complexity of Government
No ratings yet
Beyond Beyond Policy Analysis: Emergent Policy and The Complexity of Government
7 pages
Sudhir Krishnaswamy - Jurisprudence-I Course Plan 2005
No ratings yet
Sudhir Krishnaswamy - Jurisprudence-I Course Plan 2005
15 pages
American Business, Public Policy, Case-Studies, and Political Theory
100% (1)
American Business, Public Policy, Case-Studies, and Political Theory
29 pages
Hadkins Helen Lewis Samantha Budden Joanna Interactive 4 Stu PDF
100% (1)
Hadkins Helen Lewis Samantha Budden Joanna Interactive 4 Stu PDF
132 pages
dp96 6 PDF
No ratings yet
dp96 6 PDF
30 pages
The Growth of American Government: Governance from the Cleveland Era to the Present
From Everand
The Growth of American Government: Governance from the Cleveland Era to the Present
Ballard C. Campbell
No ratings yet
Strategies for Governing: Reinventing Public Administration for a Dangerous Century
From Everand
Strategies for Governing: Reinventing Public Administration for a Dangerous Century
Alasdair Roberts
No ratings yet
The Quality of Democracy: Theory and Applications
From Everand
The Quality of Democracy: Theory and Applications
Guillermo O'Donnell
No ratings yet
Comparing Public Bureaucracies: Problems of Theory and Method
From Everand
Comparing Public Bureaucracies: Problems of Theory and Method
B. Guy Peters
No ratings yet

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.